JPH0330557B2 - - Google Patents

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to an automatic steering system for ships, and more specifically, when a ship navigates under disturbances such as waves and wind, the propulsion energy is This invention relates to an energy-saving automatic steering system for a ship that uses the rate of increase from propulsion energy as an evaluation function, and steers the ship so that the evaluation function value becomes the minimum, and is adaptive to dynamic characteristics and disturbances.

[Conventional technology]

As is well known, a conventional automatic steering system for a ship is intended to cause a ship to sail straight ahead in a set direction, that is, to minimize deviation in direction as much as possible. For this purpose, feedback control is performed to steer the vehicle by performing calculations such as proportionality, differentiation, and integration on the difference between the actual orientation detected by a gyro compass and the set orientation. More recently, it has become possible to navigate in a set direction, and
Autopilot systems are emerging that control the vehicle in a way that aims to save energy. When performing control with such a device, its gain is adjusted to minimize a certain evaluation function. The evaluation functions that have been proposed and adopted so far are, for example, J ₁ = λ ₁ ² + λ ₂ ² + λ ₃ ψ〓 ² ... (1) where ² is the root mean square value of the steering angle, and ² is the azimuth. The root mean square value of the angle, ψ〓 ² is the root mean square value of the azimuthal angular velocity, λ ₁ , λ ₂ , λ ₃ are weighting coefficients, and the optimal control law is obtained by calculation using optimal regulator theory, etc. There is.

In this type of control, methods are used to identify the motion state of the ship during navigation from various detected quantities, or to vary the weighting coefficient depending on the state. The above-mentioned evaluation function value is minimized based on the azimuth angle detected in . Such an evaluation function J ₁ is expressed only by the square term as described above, and even when taking into account the oblique angle described later, the oblique angle can be approximated using the rudder angle and the azimuthal velocity. I'm looking for it. Such an evaluation function is used because it consists only of square terms, and control to minimize its value can be performed extremely easily using modern control theory that has existed for a long time. This is based on the reason that it is not easy to detect the oblique angle β with high accuracy.

[Problem to be solved by the invention]

Control using such an evaluation function has the following drawbacks. The above-mentioned evaluation function J ₁ itself means trying to keep the deviation as small as possible by using the smallest possible amount of operation.In other words, when applied to ship control, when a ship deviates from a straight-ahead state, it returns it to a straight-ahead state. Therefore, it only means that it can be returned with the minimum J ₁ . Therefore, it does not have a true energy saving function for ships that navigate under disturbances as will be described in detail later. In other words, the ship sails diagonally not only when the helm is steered, but also when it is subjected to disturbances from the side or diagonally.If the slant angle β is approximated only by the rudder angle δ or the azimuth angular velocity ψ, then the ship will sail diagonally not only when the helm is steered, but also when it receives disturbances from the side or diagonally. If oblique sailing occurs, the above approximation becomes completely unreasonable, and appropriate steering control that minimizes combustion consumption will not be performed.

For example, in the ship operation control system described in Japanese Patent Application Laid-Open No. 55-11968, the azimuth angle ψ and the oblique angle β, which are independent components of the ship's motion, are forcibly considered to be dependent, and only the azimuth angle ψ is fed back. It's in control. In other words, it is not designed to feed back both the azimuth angle ψ and the oblique angle β. Therefore, although it can be said that the device disclosed in Japanese Patent Application Laid-Open No. 11968/1983 is of the dynamic characteristic adaptive type, it is not of the disturbance adaptive type. This is due to incorrect derivation of the control law.

The present invention has been made in order to eliminate the above-mentioned drawbacks, and its purpose is to take into account the above-mentioned disturbances when the controller adjusts the rudder angle and navigates the ship toward a target course. An object of the present invention is to provide an automatic steering system for a ship that enables energy-saving navigation that keeps fuel consumption to a minimum during a voyage.

That is, by adopting the correct state variables (mutually independent azimuth angle ψ and oblique angle β), the correct evaluation function J ₂
(non-positive definite integrand) is used to process irregular disturbances as if they were based on reality. Furthermore, it is intended to realize energy saving in the true sense of the word by employing correct optimal control and dynamic characteristics that can be optimized according to changes in disturbance, and by being a disturbance adaptive type.

Here, the above-mentioned state variables will be explained below. The hull motion is expressed by the azimuth angle ψ and the drift angle (crossflow angle) β. These are caused by being exposed to irregular disturbances such as waves and wind, and by steering.

ψ=ψ _d +ψ _r β=β _d +β _r ψ _d , β _d is ψ due to disturbance, β ψ _r , β _r is ψ due to rudder, β (a) These ψ _r and β _r are the input rudder angle δ This is the output of

ψ〓 _r =F ₁ δ β _r =F ₂ δ ψ〓 _r = d/dtψ _r (d/dt is time derivative) F ₁ and F ₂ are operators (b) Approximately, β _r = c ₁ ψ 〓 _r + c ₂ δ c ₁ , c ₂ can be expressed as a constant (c).

However, ψ _d and β _d are independent. This is because it is caused by irregular disturbances whose components are mutually independent turning moment and lateral force.

If ψ _d and β _d are independent, ψ and β are independent. Therefore, in order to find δ that minimizes the evaluation function (evaluation criterion) composed of ψ, β, and δ, it is necessary to feedback not only ψ but also β.

However, in the above-mentioned Japanese Patent Application Laid-Open No. 55-11968, β=c ₁ ψ〓+c ₂ δ c ₁ and c ₂ are constants (d), and only ψ is fed back.
For the reasons mentioned above, formula (d) is incorrect.

In addition, in the present invention, we were able to realize this because we found a new evaluation function that enables energy saving and also became able to accurately detect the oblique angle, but the latter is beyond the scope of this paper. Since this is not the gist of the invention, the explanation will be omitted.
The details of the former new evaluation function will be described later, but the outline is as follows. As mentioned above, the conventional evaluation function _J1 aims to reduce the azimuth angle, but even if a ship navigates to a certain target with a large azimuth angle, there is no problem as long as fuel consumption is minimized. do not have. The present invention is based on the finding that the increase in propulsion energy can be kept to a minimum by minimizing the evaluation function employed therein.

Note that the integrand of the evaluation function is ψ, ψ〓, β,
It is a function of δ and is not positive definite (ψ〓=d/dtψ). In JP-A-55-11968, the general form is ψ,
Although it is expressed as a function of ψ, β, δ [Equation (8)′ in the publication], what is used is a function of ψ, ψ〓, δ [Equation (8) in the publication], This is positive definite.

By the way, irregular disturbances are colored noise. However, in Japanese Patent Application Laid-open No. 11968/1983, white noise is assumed based on the process of deriving the optimal control law. Regarding the optimal control law, Japanese Patent Laid-Open No. 11968/1983 applies the optimal regulator theory assuming that the integrand of the evaluation function is positive definite and that the disturbance is white noise. This is not true.

[Means to solve the problem]

The present invention estimates the strength of disturbances caused by waves, wind, etc. by taking into account the oblique angle of the hull 2 in a certain state during navigation, when adjusting the rudder angle using the controller 4 and navigating the ship along the target course. The gain of the controller 4 when minimizing the evaluation function value regarding fuel consumption is estimated from the strength of this disturbance, and the controller 4 having this gain is used as an automatic steering device to navigate the ship toward the target. be. and a dynamic characteristic calculating means for making the dynamic characteristic adaptive type, a disturbance calculating means for making the disturbance adaptive type,
It has a gain output means for determining the correct optimum gain, and uses independent azimuth angle ψ and oblique angle β as state variables in order to correctly represent the system, realizing energy saving in the true sense.

The feature of the automatic ship steering system of the present invention is that, with reference to FIG. A dynamic characteristic calculating means 5 for estimating, a disturbance calculating means 6 for estimating the strength of the disturbance from the detected azimuth ψ and oblique angle β, and a disturbance output from the disturbance calculating means 6 and the dynamic characteristic calculating means 5. A sailing simulation section 10 is provided which simulates the sailing state of a ship controlled by the controller 4 using the strength of the ship and the dynamic characteristics of the ship's body 2. An evaluation function calculation unit 11 calculates an evaluation function value based on the rudder angle δ, azimuth angle ψ, azimuth angular velocity ψ〓, and oblique angle β output from the navigation simulation unit 10.
and an optimal gain search unit 12 that searches for the gain of the controller 13 of the navigation simulation unit 10 that minimizes this evaluation function value, and a gain output means 7 that outputs the optimal gain to the controller 4. There is.

〔Effect of the invention〕

According to the present invention, when the controller adjusts the rudder angle and navigates the ship along the target course, disturbances that act on the ship while sailing at a set ship speed are taken into consideration, and such disturbances and the state of the ship are maintained. As long as fuel consumption is kept to a minimum, true energy saving operation of the vessel can be achieved. Therefore, whenever there is a change in external disturbances, energy-saving navigation is performed in accordance with the change, and the ship can be operated with the lowest fuel consumption throughout the entire voyage to the final destination. In other words, it is not only possible to optimize according to changes in dynamic characteristics;
It can be optimized in response to changes in irregular disturbances such as waves and wind, achieving true energy savings.

〔Example〕

Hereinafter, the present invention will be explained in detail based on examples thereof.

FIG. 1 is an overall system diagram of a ship automatic steering system 1 according to an embodiment of the present invention. This includes a controller 4 that commands the rudder angle to the rudder 3 of the hull 2, a dynamic characteristic calculation means 5 that estimates a motion differential coefficient representing the dynamic characteristics of the hull 2, and a disturbance calculation means 6 that estimates the strength of the disturbance.
Its main components are: and a gain output means 7 for outputting an optimum gain.

More specifically, the dynamic characteristic calculating means 5 calculates the rudder angle δ, azimuth angle ψ, and oblique angle β detected at certain time intervals from when the ship changes speed under certain conditions while the ship is sailing. and the azimuth angular velocity ψ〓 obtained from the azimuth angle ψ, the dynamic characteristics determined by the ship speed, cargo, and remaining fuel status of the hull 2 at that time are estimated, and the above values are given to the equation of motion of the hull 2. By doing so, a motion differential coefficient at that point in time is calculated, and this coefficient is outputted to a navigation simulation section 10 of the gain output means 7, which will be described later.

The disturbance calculation means 6 calculates the azimuth angle ψ similar to that described above.
The strength of the disturbance caused by waves, wind, etc. at that time is calculated from the oblique angle β and the obtained azimuth angular velocity ψ〓, and the strength of the disturbance is output to the navigation simulation section 10. Note that the strength of this disturbance is the turning moment and side force generated on the hull 2 by waves and wind, and even if these are subjected to spectrum processing in the frequency domain, they are also processed in the time domain. However, the figure shows the former. In this case, a hull motion spectral density calculation section 8 and a disturbance spectral density calculation section 9 are provided, and the former calculates each motion spectral density in the frequency domain from the azimuth angle ψ, oblique angle β, and azimuth angular velocity ψ〓. Based on these spectral densities, the latter calculates the disturbance spectral density in the frequency domain. These disturbance spectral densities are then output to the navigation simulation section 10, and used together with the above-mentioned motion differential coefficients to simulate navigation conditions in various types of virtual steering.

The gain output means 7 includes a navigation simulation section 10 for simulating navigation conditions, an evaluation function calculation section 11, and an optimum gain search section 12, and outputs the optimum gain to the controller 4. The navigation simulation part 10
is calculated using a mathematical model of the controller 13, the rudder 14, and the hull 15, and uses the motion differential coefficient of the hull 2 and the disturbance spectrum density described above to calculate various types of rudder angle adjustment signals by the controller 13. That is, a number of simulations are performed for different values of the angular gain. Then, in each simulation, the rudder angle δ _s , azimuth angle ψ _s , and oblique angle β _s
and the azimuth angular velocity ψ〓 _s determined from the azimuth angle ψ _s , or the respective spectral densities and mutual spectral densities are output to the evaluation function calculation unit 11 .

Here, a new evaluation function employed in the present invention will be explained. The new evaluation function J ₂ is given by the following equation (6), which indicates the rate of increase in propulsion energy taking into account the oblique angle due to disturbance. This evaluation function J ₂ is derived from the following logic.

As shown in Figure 2, if the coordinate origin of the hull 2 is set at its center of gravity G, the propulsion energy E required to make one voyage while meandering or deviating from the course under disturbance is E=∫〓 ₀ It is expressed as Tudt... (2). Note that T is the thrust, u is the forward speed of the hull 2, that is, the oblique angle β cosine of the ship speed U, and τ is the time required for one voyage. By the way, the propulsion energy E ₀ when the ship moves straight without steering is expressed as E ₀ =∫〓 ⁰ ₀ T ₀ u ₀ dt...(3), and the ship 2 does not meander or sail diagonally, This is the propulsion energy when ideally sailing in the bow direction while keeping the rudder angle at zero. Note that τ ₀ is generally τ
smaller. From these equations (2) and (3), the propulsion energy increase rate J ₂ is expressed as J ₂ = (E-E ₀ )/E ₀ ... (4). The propulsion energy required when navigating under disturbances such as
This is the rate of increase compared to that required when traveling straight on flat water. If we rearrange this equation (4) by considering the equation of motion of the hull 2 in the forward direction, we get J ₂ = μ ₁ (−) ² + μ ₂ (β + γψ〓) ψ〓 + μ ₃ ² + μ ₄ ² …… (5) Become. Here, the - symbol represents the time average value,
For example, δe is expressed as ² = (1/τ ₀ )∫〓 ⁰ ₀ δ ² _e dτ. In addition, μ ₁ , μ ₂ , μ ₃ , μ ₄
is a constant, and γ is a constant representing distance. Also, δe
is the effective steering angle, and with a and b as constants, δe = δ−
It can be expressed as aβ−bψ〓. In addition, the above formula
The physical meaning of each term in (5) is as follows. The first term (-) ² is the loss of propulsion energy due to the ship meandering, which makes the route longer than when the ship is going straight. Loss of propulsion energy due to resistance. Item ^3.2 refers to the loss of propulsion energy due to resistance caused by the ship sailing diagonally. Item ^4.2 refers to the loss of propulsion energy due to steering operation. represents loss.

Further rearranging the above equation (5), the evaluation function J ₂
is expressed by the following formula.

J ₂ =λ ₁ ² +λ ₂ ² +λ ₃ βψ〓＋λ ₄ ψ〓 ² +λ ₅ ² +λ ₆ +λ ₇ +λ ₈ δψ〓……(6) Note that δ is the rudder angle, β is the oblique angle, and ψ is the azimuth. angle, ψ〓
is the azimuthal velocity, and λ ₁ to λ ₈ are constants.

The remaining configuration of the gain output means 7 described above will be described again. The optimal gain search unit 12 searches for the gain of the controller 13 using a method such as the hill climbing method, for example, so as to minimize the evaluation function value calculated by the evaluation function calculation unit 11. This is output as the optimum gain to the controller 4 that adjusts the steering of the hull 2. Note that, for gain input to the controller 4, an automatic optimum control gain setting section 16 is provided inside or outside the gain output means 7 as necessary.

According to such an embodiment, the ship can be automatically steered so as to minimize fuel consumption in the following manner.

In Figure 1, first, the course ψ ₀ toward the target
is input to the controller 4. Now, for example, if PID control is performed in the controller 4, each gain of the PID input in advance
A steering command is issued using Kp ₀ , Ki ₀ , and Kd ₀ so that the rudder 3 has a steering angle δ ₀ . As a result, the hull 2 has a rudder angle δ ₀
Accordingly, the ship takes the target course ψ ₀ and navigates. However, if disturbances such as waves and wind act on the hull 2 during such navigation, the hull 2 will be damaged as shown in Figure 2.
takes a course ψb that deviates from the course ψ ₀ by an azimuth ψ,
Further, in some cases, sideslip may occur, resulting in a left-right velocity v, resulting in a slanting angle β. Incidentally, although the disturbance acting on the hull 2 may change suddenly, it usually lasts for a certain period of time, so while the disturbance continues, the ship is sailing at the set speed U while heading towards a certain target. fuel consumption must be minimized. In order to calculate the amount of steering for that purpose, the above azimuth angle ψ and oblique angle β
is detected by a measuring device (not shown), and the detected amount and the steering angle δ ₀ are input to the dynamic characteristic calculation means 5. Note that, as mentioned above, the dynamic characteristics of the hull 2 at that time vary depending on the ship speed, the position of the center of gravity, etc.
When a different ship speed is set during navigation, or when a situation arises in which the displacement or center of gravity differs, the above-mentioned detection is performed each time, and the dynamic characteristics in that state are calculated.

The above-mentioned azimuth angle ψ, slanting angle β, and the rudder angle δ ₀ at that time are detected, and the azimuth angular velocity ψ is calculated from the azimuth angle ψ. A motion differential coefficient that indicates the dynamic characteristics of the hull 2 at
0 and treated as the dynamic characteristics of the hull 15.

On the other hand, the azimuth angle ψ and oblique angle β are similarly inputted to the disturbance calculating means 6 at regular intervals of, for example, several minutes to ten-odd minutes, and the azimuth angular velocity ψ〓 is calculated. Strength is calculated. The process begins with the hull motion spectral density calculation unit 8 calculating the motion spectral densities Φββ, Φψψ, Φβψ〓 for the frequencies of the azimuth angle ψ, the oblique angle β, and the azimuth angular velocity ψ〓 as shown in Fig. 3 a to c. is calculated. Next, in response to this spectral density, the disturbance spectral density calculation unit 9 calculates the intensity of the disturbance with respect to the frequency,
That is, the spectral densities Φyy, Φnn, Φyn of turning moment, lateral force, etc. are estimated and calculated as shown in FIGS. 4a and 4b.

The spectral densities Φyy, Φnn, and Φyn of this disturbance are input to the gain output means 7, and a simulation of the sailing state of the hull 2 is performed based on the motion differential coefficient that has already been input to the hull 15 from the dynamic characteristic calculation means 5. , is performed in the navigation simulation section 10, which is a mathematical model. At this time, the optimal gain search section 12 searches for the gain of the controller 13 using a method such as a hill climbing method, so that the evaluation function value calculated by the evaluation function calculation section 11 is minimized. Then, this gain is input to the controller 4 that adjusts the steering of the actual hull 2 via the optimal control gain automatic setting section 16 as the optimal gain that provides the minimum fuel consumption while the hull 2 is traveling in the above-described state. That is, the aforementioned gains Kp ₀ , Ki ₀ ,
Kd ₀ is replaced by Kp ₁ , Ki ₁ , and Kd ₁ which are the optimal gains. Therefore, the controller 4 issues a steering signal based on the new gain. Hull 2
While the vessel is sailing toward a desired target while the above-mentioned disturbance intensity and the speed of the hull 2 remain unchanged, a voyage is carried out in which fuel consumption is substantially minimized.

It goes without saying that whenever the intensity of the disturbance or the state of the hull 2 changes, the series of processes such as the above-mentioned detection and calculation are performed each time. Therefore, even if the set ship speed is changed until the ship reaches its final destination, the amount of fuel consumed during the voyage to the target can be minimized, and overall, the amount of fuel consumed during one voyage can be minimized. Consumption can be reduced.

To summarize the above briefly, when the rudder angle is adjusted by the controller and the ship is sailing along the target course, the strength of disturbances caused by waves, wind, etc. is estimated by taking into account the oblique angle of the ship in a certain state during navigation. Then, from the strength of this disturbance, a gain is estimated to minimize the evaluation function value related to fuel consumption, and this gain is used to operate the controller and navigate the ship toward the target with the minimum fuel consumption. can.

As can be seen from the above description, when the controller adjusts the rudder angle and navigates the ship along the target course, the present invention also takes into account disturbances that act on the ship during navigation at a set ship speed, and eliminates such disturbances and As long as the condition of the ship's hull is maintained, fuel consumption can be kept to a minimum, allowing the ship to operate in true energy-saving mode. Therefore, whenever there is a change in external disturbances, energy-saving navigation is performed in accordance with the change, and the ship can be operated with the lowest fuel consumption throughout the entire voyage to the final destination.

[Brief explanation of the drawing]

Fig. 1 is an overall system diagram of an automatic steering system for a ship which is an embodiment of the present invention, Fig. 2 is a dynamic model diagram of a ship body undergoing disturbance, Figs. 3 a to c are motion spectrum diagrams, and Fig. 4 a and b are disturbance spectrum diagrams. 2...Hull, 4...Controller, 5...Dynamic characteristic calculation means, 6...Disturbance calculation means, 7...Gain output means, 10...Navigation simulation section, 1
1...Evaluation function calculation unit, 12...Optimum gain search unit, 13...Mathematical model controller, δ, δ ₀
... Rudder angle, β ... Oblique angle, ψ ... Azimuth angle.

Claims (1)

[Scope of Claims] 1. In an automatic steering system that constitutes a closed loop that steers the ship according to the difference between the target heading and the actual heading, the ship's hull at that time is determined based on the rudder angle, azimuth angle, and oblique angle detected during navigation. a dynamic characteristic calculation means for estimating the dynamic characteristic of the disturbance; a disturbance calculation means for estimating the strength of the disturbance from the detected azimuth and oblique angle; A sailing simulation part that simulates the sailing state of the closed-loop ship using the strength and dynamic characteristics of the ship, and is expressed as a function of the rudder angle, azimuth angle, directional angular velocity, and oblique angle output from this sailing simulation part. It has an evaluation function calculation section that calculates an evaluation function value representing a propulsion energy increase rate, and an optimal gain search section that searches for a gain of the controller of the navigation simulation section that minimizes this evaluation function value, and constitutes the closed loop. An automatic steering device for a ship, comprising: gain output means for outputting the optimum gain to a controller;