Over the years, ocean racing yachts have continued to evolve in terms of architecture, choice of materials and profiles to gain ever greater performance. Amongst these latest advances, electronics and computers have become major points in order to optimise the tuning of these boats and assist the skipper in steering them.
Let's take a look at the electronics on board (Figure 1). On a racing yacht, dozens of data points (wind, boat speed, etc.) are measured by sensors. All of this data is collected, processed and stored by a navigation centre (Bravo 4). A large number of sensors can be linked to the control unit thanks to MiniNodes, which are small acquisition modules that transmit the sensor information. These are small acquisition modules that transmit sensor information. They increase the connection capacity of the navigation centre tenfold. The data processed by the control unit is then used either to analyse the boat's performance or is transmitted to the various navigation aid systems (routing software, tactical analysis, display, etc.) and servo systems (MAD Controller DCMotor), in particular the autopilot (MAD Brain Autopilot).
Figure 1: The electronics of a servo-driven sailboat
An autopilot consists of two successive control loops. A "general" loop, thecontrol loop, which is managed by the MAD Brain. It is in this first loop that all the know-how and intelligence of the pilot resides. Its role is to determine the rudder angle that will allow the boat to respect the instructions specified by the skipper (a course to follow, an apparent wind angle to respect, etc.). In order to determine the appropriate rudder angle, the steering control algorithm relies on all the information sent back by the navigation unit (speed, apparent/actual wind, etc.). This angle command is then transmitted to the MAD Controller, which is responsible for ensuring that it is respected. This is the second loop, theservo control of the actuator. The MAD Controller manages the electrical power applied to the cylinder or the hydraulic pump to operate the bar. Its role is to adjust this power according to the tiller angle setpoint transmitted by the MAD Brain but also according to the current position of the tiller directly received via the tiller sensor or via the central unit if a processing of the measurement is necessary (e.g.: case of an installation with two sensors (tricky sensor)).
All of these control loops assist the skipper in handling the boat and provide certain safety functions (anti-cavity, etc.). However, manoeuvring a rudder or a foil is energy-intensive and high levels of power must be deployed. The servo loops set up for steering and controlling the actuators must take this criterion into account, otherwise there is a risk of running out of energy on board. And without energy, the boat becomes much more difficult to manage, and it is strongly advised to return to port (watermaker shutdown, loss of geolocation data, communication, obligation to steer 24 hours a day, etc.).
The purpose of this article is to explain how Madintec integrates the energy dimension in parallel with performance in its developments. It describes the entire process of sizing systems, from actuators to electrical storage, allowing for the best possible management of the energy criterion in a servo system. This work enables Madintec to control consumption due to steering (bar or foil) as well as to improve the precision of the servo systems thanks to a better knowledge of the power system.
Electronics and energy
To illustrate this, let's take the example of the servo loop of an actuator, the control of the "cant" of a foil using an electric cylinder.
Before we can talk about electrical power, it is necessary to determine the mechanical powers involved. Indeed, the electrical power is used above all to provoke a mechanical effort allowing the foil to be manoeuvred. These measurements or estimates of effort are carried out by the team's architects or design offices using modelling tools or by direct instrumentation of the boat. The forces/loads applied to the foil are then recovered to design the actuators.
In order to be able to deliver the force required to move the foil, it is necessary to correctly size the cylinder that will operate it. It is important to break down all the elements that make up the actuator in order to understand its behaviour. An electric actuator is composed of a DC motor that transforms electrical energy into mechanical energy, a reducer, a screw-nut system and a rod that transform and restore the mechanical forces generated by the motor. Through modelling, it is possible to completely control the behaviour of each of these elements, particularly in terms of energy loss. Indeed, the energy transmitted to the motor will not be entirely returned to the foil, we speak of energy efficiency. A DC motor will suffer electrical and magnetic losses which result in heating (Joule losses and iron losses), as well as mechanical losses due to rotor friction. The gearbox, the screw-nut system and the spindle are also subject to friction and mechanical backlash, which results in energy losses (Figure 2). Knowing the mechanical power to be applied to the foil, the behaviour of each of the elements making up the cylinder (and therefore its efficiency), it is then possible to determine the electrical power to be applied to the motor in order for it to be able to restore the necessary mechanical force.
Figure 2: Electric cylinder
However, knowledge of power does not yet allow us to talk about energy. Energy corresponds to the application of power over a certain period of time. It is therefore necessary to know how the electric cylinder will be used. It is the role of the MAD Controller to control the cylinder and thus to manage this servoing (respect of an angle set point). It is then important to set up adapted control strategies in order to guarantee the precision of the system while reducing the energy consuming character of the movement of a foil. Some control strategies will be detailed in the following.
The knowledge of these strategies makes it possible to know the way in which the cylinder will be solicited and thus the energy consumption necessary for the servoing of a foil. In order to be able to supply this energy to the system, it is necessary to be able to store it. It is therefore important to size the energy storage correctly. The sizing of the storage and the choice of a suitable technology is done after reviewing the following points:
- Frequency study of the storage system (How often is the energy storage system used from the point of view of both energy consumption and production?)
- Estimation of current draw (How much is it being drawn? What powers are involved?)
- Determination of the useful capacity (How much energy is needed? How much energy will the storage system be able to deliver without impairment? What are the safety margins?)
- Definition of constraints (weight, size, financial and environmental costs)
Once all these points have been taken into account, it is then possible to select a suitable storage technology (diesel, methanol, Li-ion battery, Pb, super-capacity, etc.). Each technology has a specific use (Figure 3). For example, diesel fuel, which has a very high energy content (around 1 kWh/kg), will be used for long-term storage. It can carry a large amount of energy for a moderate weight. However, it does not withstand too frequent use. It is not advisable to start a generator every two minutes, as the generator has a very low efficiency at start-up and the energy losses would be too high. This is why it is often customary to transfer this energy to lithium-ion batteries, which have a lower mass energy (about 100Wh/kg) but which can withstand much more frequent use. They are used for medium-term storage. At the extreme end of the spectrum are super-capacitors, which generate much more power than the two previous technologies (around 5kW/kg compared with 200W/kg for Li-ion batteries) but which will be used for short-term storage because they are very heavy in energy terms (around 10Wh/kg). This technology is used, for example, to move the keel of a boat.
Figure 3: Ragone diagram (source: ac-nantes.fr)
All of these steps allow us to know the energy requirements for moving a foil, from the mechanical power to be exerted through the energy efficiency of the cylinder, the electrical powers to be supplied and the storage capacities and characteristics. Once this behaviour is known, what energy-related control strategies could be put in place?
Energy" strategies
Madintec, among other developments, is working on implementing the following strategies in the control of its actuators. The team is committed to reducing equipment consumption while ensuring at least the same level of accuracy as current systems.
- Optimising engine efficiency
An electric motor has an energy efficiency that can vary from 30% to 80% depending on its different operating points. A significant part of the energy can therefore be lost if the commands transmitted to the motor are not well defined in relation to the load to be moved. The control strategies applied by Madintec seek to maximise this motor efficiency by selecting appropriate accelerations and travel speeds. It is therefore not simply a question of moving a foil from point A to point B but of finding the best path for the motor to be as efficient as possible. Controlling the system allows us to quantify the energy we want to allocate to the manoeuvre.
- The implementation of a "weighted" control
The two main criteria in setting up control strategies are performance and energy consumption. Madintec expresses these two criteria using cost functions representative of these criteria. The controller generates commands that will allow to minimize or maximize these functions (minimize consumption/maximize performance). A weighting can be set up to favour one or other of these criteria. A balance of these weightings can then be made according to the skipper's objectives. It is also possible to base this on the boat's current energy balance. We can also imagine basing ourselves on a balance sheet within 24 hours by estimating energy production and consumption, knowing the boat's approximate route. This work is still in progress, but we are obtaining very encouraging results.
In order to implement these strategies, Madintec relies on the theory of optimal control which is part of the science of automation. This type of control is widely used in aeronautics. It goes further than the simple PID (Proportional Integral Derivative) corrector currently used by the old generation of pilots and also found in production systems. This type of corrector was no longer sufficient to meet the new needs of ocean racing in terms of increasingly demanding servo-control (foil control, 3D control, multi-criteria control (performance, energy)).
Currently, the Madintec team is continuing its research and development work to integrate the energy dimension as well as possible while guaranteeing (or even improving) the same level of performance as the state of the art. The energy consumption of an autopilot represents a large part of the total consumption of a boat. By working solely on the actuator servo loop (power control of the electric cylinder), Madintec has achieved significant gains in terms of energy consumption. These results open up good prospects for our future developments. Madintec is now continuing its work by no longer basing itself solely on the actuator's servo loop but on the complete servo control. More in-depth work on the algorithms will make it possible to eliminate as much as possible the parasitic movements of the driver and thus further reduce energy consumption.
In a few years' time, perhaps we will be able to move on to sailing ships that run solely on green energy. Today's sailing ships are already sufficiently instrumented. By controlling consumption as much as possible, we will be able to make more use of this network and obtain sailing boats that are autonomous in energy.
