Prediction method for dynamic torque of propulsion load of propeller

Abstract

The invention relates to the technical field of ship power equipment, in particular to a method for predicting the dynamic torque of a propeller propulsion load. Comprising the steps of constructing a mathematical relationship between propeller rotating speed and torque; predicting the rotating speed of the propeller based on an ANFIS algorithm; propulsive load dynamic torque prediction. The method can accurately predict the dynamic torque of the propulsion load, helps the ship to realize the optimal performance, and reduces the fuel consumption to the greatest extent, thereby reducing the operation cost and improving the economic benefit.

Description

Technical Field

[0001] The invention relates to the technical field of ship power equipment, and in particular to a method for predicting the dynamic torque of a propeller propulsion load. Background Art

[0002] In the power system of a ship, the propeller, as a propulsion device, undertakes the core task of converting power into propulsion. By accurately predicting and controlling the propeller torque, the ship can optimize the sailing plan and speed control under different sailing conditions, thereby minimizing fuel consumption and reducing operating costs. In addition, accurate prediction of propeller torque is also of great significance to ship design and route planning, and can optimize the propulsion system to reduce fuel consumption and emissions. Reasonable control of propeller torque can also reduce the load of the power system, extend the service life of the equipment, and reduce maintenance costs. Therefore, effective prediction and control of propeller torque is not only beneficial to the economic benefits of the ship, but also helps to reduce the negative impact on the environment, which is in line with the trend of sustainable development in the shipping industry and brings dual economic and environmental benefits to ship operators. Summary of the invention

[0003] In view of the above problems, the present invention proposes a method for predicting the dynamic torque of a propeller propulsion load, which specifically includes:

[0004] (1) Construct a mathematical relationship between propeller speed and torque;

[0005] (2) Predict propeller speed based on ANFIS algorithm;

[0006] (3) Prediction of propulsion load dynamic torque.

[0007] Preferably, constructing a mathematical relationship between propeller speed and torque includes:

[0008] Establish the propeller speed n prop_i With torque M prop_i The exponential relationship between the outputs; as shown below,

[0009]

[0010] Where α i ,β i are the fitting coefficient and index of the propeller, respectively;

[0011] The exponential relationship is derived into a first-order linear mathematical relationship by natural logarithm, as shown in the following formula:

[0012] Prioritize the propeller speed prediction based on the ANFIS algorithm, specifically including:

[0013] Use an exponential membership function to obtain fuzzy clustering input values:

[0014]

[0015] Where μ AiorBi is the membership function; n prop_i , Δn prop_i are propeller speed and prediction deviation respectively; a i 、c i is the membership related parameter;

[0016] For the membership function μ AiorBi Create weights; as shown below,

[0017]

[0018] The weights are standardized as shown in the following formula:

[0019]

[0020] Defuzzification is achieved by calculating the weight of each node. The calculation formula is as follows:

[0021]

[0022] Where p i ,q i and r i There are three conclusion parameters respectively. As a rule, the number of conclusion parameters of each rule should be greater than the number of input parameters;

[0023] The propeller prediction value is obtained as follows:

[0024]

[0025] Where n ANFIS is the propeller prediction value obtained by the ANFIS algorithm.

[0026] Preferably, the propulsion load dynamic torque prediction specifically includes: constructing a propulsion torque dynamic estimation model, and the propulsion torque dynamic estimation model includes an offline learning process and an online learning process.

[0027] The beneficial effect of the present application is that it accurately predicts the dynamic torque of the propulsion load, helps the ship achieve optimal performance, minimizes fuel consumption, thereby reducing operating costs and improving economic benefits. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the specific implementation of the present invention or the technical solution in the prior art, the following briefly introduces the drawings required for use in the specific implementation or the description of the prior art.

[0029] Figure 1 A flow chart of a method for predicting dynamic torque of propeller propulsion load provided in accordance with an embodiment of the present invention;

[0030] Figure 2 A flowchart of a propeller speed prediction method based on ANFIS is provided for one embodiment of the present invention. DETAILED DESCRIPTION

[0031] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0032] Reference Figure 1 and Figure 2 , as an embodiment of the present invention, provides a method for predicting dynamic torque of propeller propulsion load, comprising:

[0033] S101: Construct the mathematical relationship between propeller speed and torque;

[0034] Propeller speed n prop_i With torque M prop_i There is a certain exponential relationship between the outputs, as shown in formula (1).

[0035]

[0036] Where α i ,β i are the fitting coefficient and exponent of the propeller, respectively.

[0037] The exponential relationship is derived into a first-order linear mathematical relationship by natural logarithm, as shown in formula (2).

[0038]

[0039] S102: Propeller speed prediction based on ANFIS algorithm;

[0040] ANFIS is used to predict the changing law of speed during operation and provide reference input values ​​for dynamic torque prediction. It is divided into five layers, such as Figure 2 The details include as follows:

[0041] (1) Fuzzification

[0042] This layer uses the exponential membership function to obtain the fuzzy clustering input value. The function expression is shown in formula (3).

[0043]

[0044] Where μ AiorBi is the membership function; n prop_i , Δnprop_i are propeller speed and prediction deviation respectively; a i 、c i is the membership related parameter.

[0045] (2) Regularization

[0046] Regularization uses μ calculated by the fuzzification layer AiorBi Create weights as shown in formula (4).

[0047]

[0048] (3) Standardization

[0049] Normalization calculates the standardized weight of each rule, where the standardized weight is the percentage of the shooting intensity of a rule to the total number of all shooting intensities, and the calculation formula is as follows:

[0050]

[0051] (4) Defuzzification

[0052] The defuzzification layer is implemented by calculating the weight of the rule calculation in each node, as shown in formula (6).

[0053]

[0054] Where p i ,q i and r i There are three conclusion parameters respectively. As a rule, the number of conclusion parameters of each rule should be greater than the number of input parameters.

[0055] (5) Sum

[0056] The prediction value of the last layer is obtained by accumulating all the outputs obtained by each rule in the fourth layer, as shown in Equation (7).

[0057]

[0058] Where n ANFIS is the propeller prediction value obtained by the ANFIS algorithm.

[0059] S103: prediction of dynamic torque of propulsion load, including:

[0060] A propulsion torque dynamic estimation model is constructed, and the propulsion torque dynamic estimation model includes an offline learning process and an online learning process.

[0061] (1) Offline learning process

[0062] The propeller dynamic process is described by considering the power (integral) and speed deviation (distance) parameters as shown in Equations (8)-(9), which are defined as the integral learning parameters M I_i and distance learning parameter M D_i .

[0063]

[0064] Where n low 、n high The minimum and maximum operating speed of the propeller respectively.

[0065] (2) Online learning process

[0066] The prediction data from ANFIS is introduced and combined into a new coupling matrix as follows:

[0067]

[0068] The formula of the offline learning process is used to further describe the online learning process as shown in equations (11)-(12).

[0069]

[0070] The probability distribution function is used to define the minimum and maximum operating speed values, as shown in formula (13).

[0071]

[0072] Where μ and σ are the mean and standard deviation of the engine speed data points, respectively.

[0073] The integral learning parameters and distance learning parameters are random variables during the calculation process, and are approximated using Gaussian kernel density estimation and superimposed Gaussian distribution, as shown in equations (14)-(15).

[0074]

[0075] Finally, the available integral parameters are determined by equation (16).

[0076]

[0077] Equation (17) is used to estimate the propeller fitting coefficients and exponents.

[0078]

[0079] The predicted value of propulsion load dynamic torque calculation is shown in formula (18).

[0080]

Claims

1. A method for predicting dynamic torque of propeller propulsion load, characterized in that: include: Construct a mathematical relationship between propeller speed and torque; Predict propeller speed based on ANFIS algorithm; Propulsion load dynamic torque prediction.

2. A method for predicting dynamic torque of propeller propulsion load according to claim 1, characterized in that: The construction of the mathematical relationship between the propeller speed and the torque specifically includes: Establish the propeller speed n prop_i With torque M prop_i The exponential relationship between the outputs; as shown in the following formula, Where α i ,β i are the fitting coefficient and index of the propeller, respectively; The exponential relationship is derived into a first-order linear mathematical relationship by natural logarithm, as shown in the following formula:

3. The method for predicting dynamic torque of propeller propulsion load according to claim 1, characterized in that: The prediction of propeller speed based on the ANFIS algorithm specifically includes: Use an exponential membership function to obtain fuzzy clustering input values: Where μ AiorBi is the membership function; n prop_i , Δn prop_i are propeller speed and prediction deviation respectively; a i 、c i is the membership related parameter; For the membership function μ AiorBi Create weights; as shown below, The weights are standardized as shown in the following formula: Defuzzification is achieved by calculating the weight of each node. The calculation formula is as follows: Where p i ,q i and r i There are three conclusion parameters respectively. As a rule, the number of conclusion parameters of each rule should be greater than the number of input parameters; The propeller prediction value is obtained as follows: Where n ANFIS is the propeller prediction value obtained by the ANFIS algorithm.

4. The method for predicting dynamic torque of propeller propulsion load according to claim 1, characterized in that: The propulsion load dynamic torque prediction specifically includes: constructing a propulsion torque dynamic estimation model, and the propulsion torque dynamic estimation model includes an offline learning process and an online learning process.

Images