CN103900541B - Marine condition estimator - Google Patents

Abstract

The invention belongs to the field of marine environment monitoring and in particular relates to a marine condition estimator, which comprises a ship rolling signal acquisition device, an AR (auto regressive) model parameter identifier, a white noise simulator, a rolling angular spectrum density estimator, a sea wave spectrum density reverse calculator and a marine condition computing unit. Compared with a measuring method based on a radar wave and a radio wave reflection principle of the radar wave, the marine condition estimator is simple in structure, needs no extra hardware cost, is accurate in estimation and can meet the requirement of supporting devices of a sea ship on marine condition information.

Description

A kind of sea situation estimator

Technical field

The invention belongs to ship environment monitoring field, and in particular to a kind of sea situation estimator.

Background technology

The afloat motion attitude control of ship is very close with sea situation relation, the control of Ship Motion Attitude control system Effect has very close relationship with sea situation, how to carry out real-time estimation to sea situation in the controls, in good time adjustment control Parameter, will greatly improve Ship Motion Attitude control effect, improve the ability that ship adapts to severe sea condition.To solve the reality of sea situation When analyze, occurred using the radar emission signal and its radar signal assembled on ship by between the echo on corrugated when Between difference and speed of the ship in metres per second calculating the wave height of wave, sea situation is determined according to the wave height for calculating then, the method is possible in theory, but It is also to be subject to affecting for the speed of a ship or plane, wave celerity etc., and its cost is very high, needs to be equipped with radar signal and its reflected signal Reception device.

The content of the invention

It is an object of the invention to provide a kind of directly obtain ship rolling letter using navigation system or angular transducer etc. Number, the sea situation estimator of sea situation is realized by the statistics to ship rolling signal.

The object of the present invention is achieved like this：

Sea situation estimator, including ship rolling signal picker, AR identification of Model Parameters devices, white noise acoustical simulation, rolling Angular spectrum density estimation device, wave spectrum density backstepping device, sea situation computing unit, ship rolling signal picker are adopted using digital A/D Collection interface realizes that the interface is responsible for the collection of ship rolling angle signal, including the Filtering Processing of collection signal；White noise acoustical simulation Using pseudo-random sequence and digital form, the white noise sequence needed in AR identification of Model Parameters is simulated, joined as AR models The stochastic inputs signal of number identifier；AR identification of Model Parameters device is used for defeated by white noise and ship rolling angle stochastic signal Enter, AR model parameters are recognized, acquisition can carry out the forecast mould of ship rolling angle signal accurate forecast using white noise Type；Roll angle spectral density function estimator, using permanent white noise spectrum density and the AR models of foundation, estimation obtains rolling Angular spectrum density function；Wave spectrum density backstepping device utilizes the transmission function relation between ship rolling angle and wave, using acquisition The anti-spectral density function for releasing wave of roll angle spectral density function；Sea situation computing unit utilizes the anti-wave spectrum density for pushing away acquisition Function, by the mean square deviation of ocean wave spectrum Density functional calculations wave, is then had between adopted wave height and wave root-mean-square by wave Relation is further calculated the estimation for having adopted wave height, completing sea situation of wave.

AR identifiers are with N number of ship rolling angular data φ for measuring_t(t=1,2 ..., N), sets up P (P ＜ N) Rank AR (P) model：

Wherein w_tIt is that zero-mean, variance areStationary white noise sequence；a_k(k=1,2 ... P) are autoregressive coefficient, can By based on any two roll angle time serieses φ in AR (P) model_k,φ_jBetween covariance r_φ(k, j) is with satisfaction Yule-Walker equations

Select correlation method, the covariance method, repair for condition One kind in the positive covariance method, to r_φ(k, j) is estimated, wherein

Correlation method

The covariance method

Modified covariance method

Covariance r is determined by estimation mode_φAfter (k, j), the autoregressive coefficient a being calculated in AR (P) model_k(k= 1,2,…P)。

There is following relation between roll angle spectrum density estimator ship rolling angle and white noise

a₀It is defined as 1

Between roll angle spectral density function and white noise density spectra, relation is：

Take spectrum density S of white noise_w(ω)=b, b are permanent number, then ship rolling angular spectrum is

Wave spectrum density backstepping device

According to random theory, the Automated generalization data between ocean wave spectrum, ship rolling angular spectrum and the rolling motion of ship

S_φ(ω)=G_φh(-jω)G_φh(jω)S_h(ω)=| G_φh(jω)|²S_h(ω)

Wherein S_φ(ω) it is ship rolling angular spectrum；G_φhS () is ship rolling motion model, when ship low-angle rolling, G_φhS () is second-order linearity function, when ship wide-angle rolling, G_φhS () is nonlinear function；S_h(ω) it is ocean wave spectrum,

The downward wave of different waves meets with spectral density function and is

In the case where frequency is met with, wave spectrum density is

Sea situation computing unit carries out variance statistic calculating by the wave spectrum density for estimating to obtain

Then the relation having between adopted wave height and wave root-mean-square according to wave, calculating has adopted ripple under current sea situation Height, and sea situation classification is compareed, forecast the sea situation grade residing for current ship

The beneficial effects of the present invention is：

Compared with based on radar wave and its radio wave attenuation principle measuring method, estimator simple structure, it is not necessary to extra Hardware cost.Estimate accurate, disclosure satisfy that demand of the marine ships corollary apparatus to sea situation information.

Description of the drawings

Sea situation estimator block diagrams of the Fig. 1 based on AR models；

Fig. 2 sea situations estimate flow process.

Specific embodiment

Below in conjunction with the accompanying drawings the present invention is described further.

Sea situation estimator structured flowchart of the present invention based on AR models is as shown in Figure 1.Estimator is mainly believed by ship rolling Number harvester, AR identification of Model Parameters devices, white noise acoustical simulation, roll angle spectrum density estimator, wave spectrum density backstepping device, sea The part such as condition computing unit constitutes.

Wherein ship rolling signal picker, is realized using digital A/D acquisition interfaces, and the interface is responsible for ship rolling angle The collection of signal, including the Filtering Processing of collection signal.White noise acoustical simulation adopts pseudo-random sequence and digital form, simulates The white noise sequence needed in AR identification of Model Parameters, as the stochastic inputs signal of AR identification of Model Parameters devices.AR models are joined Number identifier is mainly used in being input into by white noise and ship rolling angle stochastic signal, AR model parameters is recognized, is obtained The forecasting model of ship rolling angle signal accurate forecast can be carried out using white noise.The estimation of roll angle spectral density function, then be Using permanent white noise spectrum density and the AR models of foundation, estimation obtains roll angle spectral density function.Wave spectrum density is anti- Drill, be, using the transmission function relation between ship rolling angle and wave, to release using the roll angle spectral density function for obtaining is counter The spectral density function of wave.It is then using the anti-wave spectral density function for pushing away acquisition, by wave spectral density function that sea situation is calculated The mean square deviation of wave is calculated, is then had the relation between adopted wave height and wave root-mean-square by wave and then is calculated having for wave Adopted wave height, and then complete the estimation of sea situation.

A.AR identifiers

In whole sea situation estimator, according to roll angle time serieses φ_k, white noise is introduced, AR (P) model parameter is carried out The AR identification of Model Parameters device principles of estimation are as follows：

If ship rolling angle is the stationary random process that a class average is zero, with the N number of ship rolling angle number for measuring According to φ_t(t=1,2 ..., N), can set up P (P ＜ N) rank AR (P) model shown in formula (1)：

Wherein w_tIt is that zero-mean, variance areStationary white noise sequence；Order P can be empirically determined, or adopts The methods such as Akaike's Information Criterion predefine an index relevant with model order P, then by optimizing index determining Order P；a_k(k=1,2 ... P) are autoregressive coefficient, can be by following based on any two roll angle in AR (P) model Time serieses φ_k,φ_jBetween covariance r_φIt is condition that (k, j) meets Yule-Walker equations (2), then can be according to reality Situation, any one of selecting type (3), (4), (5) is to r_φ(k, j) is estimated.Covariance r is being determined by estimation mode_φ(k, J), after, the autoregressive coefficient a in AR (P) model is calculated using Yule-Walker equations (2)_k(k=1,2 ... P).

1) correlation method

2) covariance method

3) modified covariance method

B. roll angle spectral density function estimates principle

Through-beam Series (1), have following relation between ship rolling angle and white noise

Formula (6) represents the transmission function shown in an accepted way of doing sth (7)：

In formula (7)

a₀It is defined as 1.

By Principle of Random Process, have between the white noise density spectra in roll angle spectral density function and formula (1) shown in formula (9) Relation

Take spectrum density S of white noise_w(ω)=b, b are permanent number, then ship rolling angular spectrum is

C. ocean wave spectrum inversion of Density principle

According to random theory, the pass having between ocean wave spectrum, ship rolling angular spectrum and the rolling motion of ship shown in formula (11) System：

S_φ(ω)=G_φh(-jω)G_φh(jω)S_h(ω)=| G_φh(jω)|²S_h(ω) (11)

Wherein S_φ(ω) it is ship rolling angular spectrum；G_φhS () is ship rolling motion model, when ship low-angle rolling, G_φhS () is second-order linearity function, when ship wide-angle rolling, G_φhS () is nonlinear function；S_h(ω) it is ocean wave spectrum.

According to formula (11), the downward wave of different waves meets with spectral density function and is

Convolution (10), in the case where frequency is met with, wave spectrum density is

D. sea situation is calculated

By the wave spectrum density for estimating to obtain, variance statistic calculating can be carried out with formula (14)

Then the relation as shown in formula (15) had between adopted wave height and wave root-mean-square according to wave, calculates current sea situation Under have an adopted wave height, and compare sea situation classification, can accurate forecast go out the sea situation grade residing for current ship.

The implementation steps of whole estimator are as follows：

The flow process that whole sea situation is estimated is as shown in Fig. 2 specific implementation step is as follows：

1). determine time period, Real-time Collection ship rolling angle, speed of a ship or plane signal；

2). the data to gathering are filtered process, pick out the wild point of measurement, filter measurement noise；

3). select AR identification of Model Parameters algorithms；

4). pseudo-random function is used, approximate white noise sequence is produced；

5). according to the white noise sequence that gathered data and computer are produced, AR models are carried out with the identification algorithm of selection The identification of parameter；

6). FFT is used, the close analysis of spectrum of white noise sequence is carried out, white noise spectrum density function is obtained, then by AR moulds Type determines roll angle spectral density function as formed filter；

7). input ship rolling parameter, the transmission function between wave and ship rolling angle, and utilize the rolling for 6) determining Angular spectrum density function is counter to push away Wave Spectrum Density Function；

8). wave variance calculating is carried out, wave variance is obtained and is had adopted wave height；

9). sea situation grade is determined according to wave height

10). it is determined that terminating no, step 1 is proceeded to if being not over), restart sea situation and estimate to calculate.

Patent of the present invention is related to the estimation problem that sea situation is carried out using Wave Information.The method proposed using this patent can So that the rolling statistical variance of ship is calculated according to the rolling motion information of ship, and the roll angle of ship is released using AR models Spectrum, it is after having rolling angular spectrum, according to the relation between STOCHASTIC CONTROL principle and ship rolling angular spectrum and ocean wave spectrum, counter to release wave Spectrum, so as to obtain the statistical parameter of sea situation, is realized the estimation to sea situation.

Claims (2)

1. a kind of sea situation estimator, including ship rolling signal picker, AR identification of Model Parameters devices, white noise acoustical simulation, horizontal stroke Cradle angle spectrum density estimator, wave spectrum density backstepping device, sea situation computing unit, it is characterised in that：Ship rolling signal picker Realized using digital A/D acquisition interfaces, the interface is responsible at the collection of ship rolling angle signal, including the filtering of collection signal Reason；White noise acoustical simulation adopts pseudo-random sequence and digital form, simulates the white noise sequence needed in AR identification of Model Parameters Row, as the stochastic inputs signal of AR identification of Model Parameters devices；AR identification of Model Parameters device is used for horizontal by white noise and ship Cradle angle signal input is recognized to AR model parameters, and acquisition can carry out ship rolling angle signal accurate forecast using white noise Forecasting model；Roll angle spectral density function estimator is obtained using the AR model assessments of permanent white noise spectrum density and foundation Obtain roll angle spectral density function；Wave spectrum density backstepping device is using the transmission function relation between ship rolling angle and wave, profit Wave spectral density function is released with the roll angle spectral density function for obtaining is counter；Sea situation computing unit utilizes the anti-ocean wave spectrum for pushing away acquisition Density function, by ocean wave spectrum Density functional calculations wave root-mean-square σ_h, then have adopted wave height h by wave_1/3With wave root-mean-square σ_hBetween relation so that be calculated wave have adopted wave height h_1/3, complete the estimation of sea situation；

Described AR identification of Model Parameters device is with N number of ship rolling angular data φ for measuring_t, wherein t=1,2 ..., N build Vertical P ranks AR (P) model：

Wherein P ＜ N,

Wherein w_tIt is that zero-mean, variance areStationary white noise sequence；a_kFor autoregressive coefficient, wherein k=1,2 ... P,

Any one of selecting type correlation method, the covariance method, modified covariance method is to r_φ(k, l) is estimated, by estimating Mode determines any two roll angle time serieses φ_k,φ_lBetween covariance r_φAfter (k, l), using Yule-Walker equations The autoregressive coefficient a being calculated in AR (P) model_k；

Yule-Walker equations：

$\left[\begin{array}{cccc}{r}_{\mathrm{\φ}}\left(0\right)& r_{\mathrm{\φ}}(-1)& ...& r_{\mathrm{\φ}}(-P)\\ r_{\mathrm{\φ}}\left(1\right)& r_{\mathrm{\φ}}\left(0\right)& ...& r_{\mathrm{\φ}}(-P+1)\\ \begin{array}{c}.\\ .\\ .\end{array}& \begin{array}{c}.\\ .\\ .\end{array}& \begin{array}{c}\\ \\ \end{array}& \begin{array}{c}.\\ .\\ .\end{array}\\ r_{\mathrm{\φ}}\left(P\right)& r_{\mathrm{\φ}}(P-1)& ...& r_{\mathrm{\φ}}\left(0\right)\end{array}\right]\left[\begin{array}{c}1\\ a_1\\ \begin{array}{c}.\\ .\\ .\end{array}\\ a_P\end{array}\right]=\left[\begin{array}{c}{\mathrm{\σ}}_w^2\\ 0\\ \begin{array}{c}.\\ .\\ .\end{array}\\ 0\end{array}\right];$

Correlation method：

${\hat{r}}_{\mathrm{\φ}}\left(k\right)=\frac{1}{N}\underset{t=0}{\overset{N-1-k}{\Σ}}{\mathrm{\φ}}_{t+k}{\mathrm{\φ}}_t,k=0,1,...,P,$

The covariance method：

${\hat{r}}_{\mathrm{\φ}}(l,k)={\hat{r}}_{\mathrm{\φ}}(k,l)=\frac{1}{N-P}\underset{t=P}{\overset{N-1}{\Σ}}{\mathrm{\φ}}_{t-l}{\mathrm{\φ}}_{t-k},l,k=0,1,...,P,$

Modified covariance method：

${\hat{r}}_{\mathrm{\φ}}(l,k)={\hat{r}}_{\mathrm{\φ}}(k,l)=\frac{1}{2(N-P)}\[\underset{t=P}{\overset{N-1}{\mathrm{\Σ}}}{\mathrm{\φ}}_{t-l}{\mathrm{\φ}}_{t-k}+\underset{t=0}{\overset{N-1-P}{\mathrm{\Σ}}}{\mathrm{\φ}}_{t+l}{\mathrm{\φ}}_{t+k}\],l,k=0,1,...,P,$

Described wave spectrum density backstepping device is according to random theory, ocean wave spectrum, roll angle spectral density function and ship rolling motion Automated generalization data between model：

S_φ(ω)=G_φh(-jω)G_φh(jω)S_h(ω)=| G_φh(jω)|²S_h(ω),

Wherein S_φ(ω) it is roll angle spectral density function；G_φh(j ω) is ship rolling motion model, when ship low-angle rolling When, G_φh(j ω) is second-order linearity function, when ship wide-angle rolling, G_φh(j ω) is nonlinear function；S_h(ω) it is wave Spectral density function,

The downward wave spectral density functions of different waves are：

$S_h\left(\mathrm{\ω}\right)=\frac{S_{\mathrm{\φ}}\left(\mathrm{\ω}\right)}{|G_{\mathrm{\φ}h}\left(j\mathrm{\ω}\right){|}^2},$

In the case where frequency is met with, wave spectral density function is：

Spectrum densities of the b for white noise, is a permanent number.

2. a kind of sea situation estimator according to claim 1, it is characterised in that：Described sea situation computing unit is obtained by counter pushing away The wave spectral density function for obtaining, carries out variance statistic calculating：

${\mathrm{\σ}}_h^2={\∫}_0^{\mathrm{\∞}}{S}_h\left(\mathrm{\ω}\right)d\mathrm{\ω}={\∫}_0^{\mathrm{\∞}}\frac{b}{\left|G_{\mathrm{\φ}h}\left(j\mathrm{\ω}\right){|}^2\right|1+\underset{k=1}{\overset{P}{\mathrm{\Σ}}}{a}_k{e}^{-j\mathrm{\ω}k}{|}^2}d\mathrm{\ω},$

Then adopted wave height h had according to wave_1/3With wave root-mean-square σ_hBetween relation, calculating has adopted ripple under current sea situation High h_1/3, and sea situation classification is compareed, forecast the sea situation grade residing for current ship；Wherein,