KR102288441B1 - Apparatus and method for estimating submarine cable status when pulling submarine cables from offshore wind power towers - Google Patents

Abstract

The present invention relates to an apparatus and a method for monitoring the condition of a submarine cable when a submarine cable is pulled out from an offshore wind power tower. To this end, the apparatus for monitoring the condition of a submarine cable includes a submarine cable state estimation module that uses a motor vector and a pulley vector as input data, and the submarine cable state data, which is data on the state of the submarine cable, as output data, and is connected to an operator client or a controller of the driving motor through a wired/wireless network, thereby capable of transmitting the submarine cable state data to the operator client or controlling the controller of the driving motor based on the submarine cable state data.

Description

Apparatus and method for estimating submarine cable status when pulling submarine cables from offshore wind power towers

The present invention relates to an apparatus and method for monitoring the condition of a submarine cable when a submarine cable is pulled out from an offshore wind power tower.

Recently, major countries are continuously promoting the transition from fossil fuel-centered energy infrastructure to renewable energy-oriented energy infrastructure such as solar power and wind power. The share of renewable energy in the OECD is increasing from 17% in 2000 to 27% in 2017 and is expected to increase to 42% by 2040, while the share of nuclear power is decreasing from 23% to 18% in the same period. The share of renewable energy in the EU was 35% of electricity generation in 2019, and wind and solar power generation (18%) surpassed coal generation (15%) for the first time. In addition, the target of the share of renewable energy is being expanded to 30% in the UK by 2030, 32% in France, and 60% in Germany by 2050.

In particular, in the case of wind power generation, floating offshore wind power has reached the level of semi-commercialization in the offshore field, and demonstrations of MW-class systems and 25-30MW-class semi-commercial power generation complexes are being built centering on Europe. The 2.3MW spar-type floating wind turbine installed off the west coast of Norway at a depth of 220 m is the first floating offshore wind turbine currently in operation.

1 is an exemplary diagram illustrating an example of the structure of an offshore wind power generation structure. As shown in Figure 1, the offshore wind power structure may be composed of an offshore structure such as an offshore wind power tower, a basic structure supporting the offshore structure. The types of foundation structures include jacket-type support structures, gravity-type support structures, monopile support structures, tri-pod support structures, semi-sub support structures, and Spar Buoys. support structures, floating support structures, and the like.

2 is an exemplary view showing an example of the structure of the monopile support structure and the jacket-type support structure of the offshore wind power generation structure. As shown in Figure 2, the monopile support structure constitutes a skirt and scour protection at the lower end of the monopile shaft to secure the bearing capacity for the ground, and the jacket-type support structure generally has four It is composed of a leg and a plurality of braces connecting between the legs, and these legs are configured to be supported on the ground by piles, and a work platform where workers can work at the bottom of an offshore structure, a foundation structure Boat landing and intermediate platforms may be constructed in the offshore part of

Republic of Korea Patent 102249783, Method for arranging the horizontal arrangement of submarine cables in the TP area of wind power tower, Taihan Electric Wire Co., Ltd. Republic of Korea Patent 10-1950711, Monitoring device for submarine cable laying machine, Dongkang M-Tech Co., Ltd.

When constructing such an offshore wind power tower, the submarine cable must be pulled up from the seabed to the offshore structure of the offshore wind power tower in order to connect a submarine cable including a power line and a communication line. 3 is a partial cross-sectional view of an offshore wind power tower using a submarine cable clamp for fixing the drawn submarine cable, and FIG. 4 shows a photograph of the jacket foundation structure of the offshore wind power tower to which the submarine cable is connected. As shown in Figures 3 and 4, the submarine cable is connected to be connected to various internal devices of power generation facilities, such as a generator, through the rising hole of the inlet of the offshore wind power tower.

When the cable is pulled into the inlet of an offshore wind power tower, if the tensile force applied to the cable is excessive, the cable may be damaged. If submarine cables are damaged, significant costs can be incurred due to cable replacement and interruption of the generation power supply. Therefore, controlling the tensile force applied to the submarine cable below a certain level (or controlling the pulling speed to adjust the tensile force) when drawing the submarine cable into the inlet of the offshore wind power tower is to connect the submarine cable of the offshore wind power tower. It is a very important task in

However, since the tensile force applied to the submarine cable is constantly changing by external forces such as currents, currents, sea winds, and waves, precise and rapid tensile force control or pull-out speed control cannot be performed.

It is an object of the present invention to provide an apparatus and method for monitoring the condition of a submarine cable when a submarine cable is pulled out from an offshore wind power tower for precise and rapid tension control or pulling speed control when a submarine cable is pulled into an inlet of an offshore wind power tower is in

Hereinafter, specific means for achieving the object of the present invention will be described.

An object of the present invention is an apparatus for monitoring the state of the submarine cable in a submarine cable pulling apparatus for drawing the submarine cable in order to draw the submarine cable in from an offshore wind power tower, wherein the submarine cable extraction device includes two pairs of A submarine cable inlet of an offshore structure comprising a plurality of pulleys composed of wheels facing each other and a driving motor providing a driving force to the wheels of the pulley, and configured on the base structure of the offshore wind power tower is configured to position the submarine cable insertion portion of the pulley, and drive motor input data or measured torque, measured speed, measured voltage or torque, speed, voltage, current or PWM duration for controlling the drive motor Driving an encoding module that uses driving motor data including driving motor measurement data for outputting the measured driving motor state including the measured current as input data, and encoding the driving motor data to use a driving motor vector as output data motor vector generation module; Using pulley rotation data, which is rotation data of each wheel constituting the pulley, as input data, encoding the pulley rotation data, and outputting a pulley vector that is a vector for rotation of each wheel configured in the pulley as output data an encoding module, a pulley vector generation module; and a submarine cable state estimation module using the driving motor vector and the pulley vector as input data, and using the submarine cable state data, which is data on the state of the submarine cable, as output data; including, a worker client or It is connected to the controller of the drive motor, and transmits the submarine cable status data to the worker client or controls the controller of the drive motor based on the submarine cable status data. This can be achieved by providing a subsea cable condition monitoring device at the time of drawing.

In addition, the submarine cable pulling device further includes a pressure sensor film configured inside the wheel to output pressure sensor data, using the pressure sensor data as input data, and encoding the pressure sensor data to obtain a pressure sensor vector It may further include a pressure sensor vector generating module using as output data, wherein the submarine cable state estimation module uses the pressure sensor vector, the driving motor vector, and the pulley vector as input data.

In addition, the submarine cable drawing device further includes a pressure sensor film installation part configured to be recessed to fit the pressure sensor film inside the wheel, and the pressure sensor film is custom-coupled to the pressure sensor film installation part. It may be characterized in that it is configured.

In addition, in the subsea cable pulling device, the pressure sensor film installation part is spaced apart by a specific distance in the direction of the rotation axis of the pulley and configured in plurality, and the pressure sensor film is configured in plurality, and is provided in each of the plurality of pressure sensor film installation parts. It may be characterized in that it is custom-coupled and configured.

In addition, the submarine cable state data may include data on whether or not the submarine cable tension overload, which is a probability value for whether the tension of the submarine cable is greater than or equal to a specific value.

In addition, the drive motor vector and the pulley vector as input data, the optimum control information generating module which uses the optimum control data, which is data for the control of the drive motor, as output data; further comprising, the driving through a wired/wireless network It may be connected to the controller of the motor to control the controller of the driving motor based on the optimal control data.

In addition, the submarine cable state estimation module is composed of a multi-layer neural network using the driving motor vector and the pulley vector as input data, and the submarine cable state data, which is data about the state of the submarine cable, as output data, , an output layer including one output node is configured to use the submarine cable state data as output data, and in a training session of the submarine cable state estimation module, the actual state data (ground truth) and the submarine cable The loss function of the submarine cable state estimating module is configured in a direction to reduce the difference in the submarine cable state data output by the state estimating module, and the submarine cable state estimating module may be updated.

In addition, the submarine cable state data includes submarine cable angle estimation data and submarine cable tension estimation data, the submarine cable state estimation module includes a first RNN cell and a second RNN cell, and the first RNN cell includes: The output data of the previous cell, RNN hidden layer information of the previous cell, and the input data of the submarine cable state estimation module are received and the submarine cable angle estimation data is output, and the second RNN cell is the output data of the previous cell. It may be configured to receive the cable angle estimation data and the RNN hidden layer information of the first RNN cell and output the submarine cable tension estimation data.

The optimal control information generating module further includes; an optimal control information generating module using the driving motor vector and the pulley vector as input data and using optimal control data, which is data for control of the driving motor, as output data; is, the optimum control information generating module is the Agent, the driving motor data of the current step and the pulley rotation data are the State, the submarine cable state data is the Environment, and the optimum is the Agent in the State in the Environment. The optimum control data output by the control information generating module to the controller of the driving motor is taken as an action, and the ratio value of the driving motor data and the actual cable pulling-out speed or the ratio value of the pulley rotation data and the actual cable pulling-out speed is A higher reward is generated as it becomes relatively large, and the optimal control information generating module, which is the agent, is configured to be updated, and connected to the controller of the driving motor through a wired/wireless network, and based on the optimal control data, the It may be characterized in that the controller is controlled.

Another object of the present invention is to provide a method for monitoring the state of the submarine cable in a submarine cable pulling device for drawing the submarine cable in order to draw the submarine cable from an offshore wind power tower, wherein the submarine cable pulling device includes two pairs Including a plurality of pulleys consisting of wheels facing each other and a driving motor providing a driving force to the wheels of the pulley, the submarine cable entry of the offshore structure configured on the foundation structure of the offshore wind power tower The drive motor input data or measured torque including torque, speed, voltage, current or PWM duration for the drive motor vector generation module to control the drive motor, Drive motor data including drive motor measurement data outputting the measured drive motor state including the measured speed, measured voltage or measured current as input data, and encoding the drive motor data to obtain a drive motor vector generating a driving motor vector outputting the output data; The pulley vector generation module takes as input data pulley rotation data, which is rotation data of each of the wheels constituting the pulley, as input data, encodes the pulley rotation data, and is a pulley vector that is a vector for rotation of each wheel configured in the pulley. a pulley vector generation step of outputting as output data; and a submarine cable state estimation step in which the submarine cable state estimation module uses the driving motor vector and the pulley vector as input data, and outputs the submarine cable state data, which is data on the state of the submarine cable, as output data; and , connected to a worker client or a controller of the drive motor through a wired/wireless network to transmit the submarine cable status data to the worker client or to control the controller of the drive motor based on the submarine cable status data , it can be achieved by providing a method of monitoring the condition of the submarine cable when the cable is pulled out from the offshore wind power tower.

As described above, according to the present invention, there are the following effects.

First, according to an embodiment of the present invention, since it is possible to control a precise and rapid tension force or draw speed control when drawing a submarine cable into the inlet of an offshore wind power tower, damage to the submarine cable is minimized when the submarine cable is pulled out is generated

Second, according to one embodiment of the present invention, since it is possible to estimate the angle of pull-in of the submarine cable when the submarine cable is pulled into the lead-in part of the offshore wind power generation tower, stable pull-out of the submarine cable is possible within a specific lead-in angle. effect occurs.

The following drawings attached to the present specification illustrate preferred embodiments of the present invention, and serve to further understand the technical spirit of the present invention together with the detailed description of the present invention, so that the present invention is limited only to the matters described in those drawings should not be interpreted as
1 is an exemplary view showing an example of the structure of an offshore wind power structure;
2 is an exemplary view showing a structural example of a monopile support structure and a jacket-type support structure among offshore wind power structures;
3 is a partial cross-sectional view of an offshore wind power tower using a submarine cable clamp for fixing the drawn submarine cable;
Figure 4 shows a photo of the jacket foundation structure of the offshore wind power tower to which the submarine cable is connected;
5 is a schematic diagram showing the configuration relationship in the offshore wind power generation tower of the submarine cable pulling apparatus according to an embodiment of the present invention;
6 is a schematic diagram showing a specific configuration of a pulley of a submarine cable pulling apparatus according to an embodiment of the present invention;
7 is a schematic diagram showing a specific configuration of a pulley of a submarine cable pulling apparatus according to a modified example of the present invention;
8 is a schematic diagram showing the operation relationship of the submarine cable condition monitoring device when the submarine cable is pulled out from the offshore wind power generation tower according to an embodiment of the present invention;
9 is a schematic diagram showing a detailed configuration of a submarine cable condition monitoring device when a submarine cable is pulled out from an offshore wind power generation tower according to an embodiment of the present invention;
10 is a schematic diagram showing a case in which the submarine cable state estimation module 13 is configured as a multi-layer neural network;
11 is a schematic diagram showing a case in which the submarine cable state estimation module 13 is composed of an attention module and an RNN encoder;
12 and 13 are exemplary views of ConvNet (CNN encoder) included in the driving motor vector generating module 10, the pulley vector generating module 11, and the pressure sensor vector generating module 12 according to an embodiment of the present invention;
14 is a schematic diagram showing a case in which the submarine cable state estimation module 13 is configured with a multi-layer RNN cell (Recurrent Neural Network cell);
15 is a schematic diagram illustrating a case in which the optimal control information generating module 14 is configured as a reinforcement learning module.

Hereinafter, with reference to the accompanying drawings, a person of ordinary skill in the art to which the present invention pertains will be described in detail an embodiment in which the present invention can be easily carried out. However, in the detailed description of the principle of operation of the preferred embodiment of the present invention, if it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted.

In addition, the same reference numerals are used throughout the drawings for parts having similar functions and functions. Throughout the specification, when it is said that a specific part is connected to another part, this includes not only a case in which it is directly connected, but also a case in which it is indirectly connected with another element interposed therebetween. In addition, the inclusion of specific components does not exclude other components unless otherwise stated, but means that other components may be further included.

Device for monitoring the condition of submarine cables when pulling out submarine cables from offshore wind power towers

Figure 5 is a schematic diagram showing the configuration relationship in the offshore wind power generation tower of the submarine cable pulling device according to an embodiment of the present invention, Figure 6 is a detailed configuration of the pulley of the submarine cable pulling device according to an embodiment of the present invention It is a schematic diagram shown. As shown in Figures 5 and 6, the submarine cable pulling apparatus according to an embodiment of the present invention, n pairs of pulleys (pulley, first pulley, second pulley) consisting of two pairs of opposite wheels (wheel) ), a pulley clamp for fixing and adjusting the length of each of the first and second pulleys, a pulley clamp shaft that is the shaft of the pulley clamp, a driving motor that provides driving force (rotational force) to a specific wheel of each pulley, and a driving motor and angle A gearbox that is connected between specific wheels of a pulley to transmit driving force, a chain that connects specific wheels of each pulley to share driving force or mutually maintains angular velocity, and is constructed on the foundation structure of an offshore wind power tower It is configured such that the space between the wheels of the lowermost pulley (submarine cable insertion part) is located in the interior of the offshore structure, in particular, in the inlet of the submarine cable.

Furthermore, with respect to the configuration of the wheels constituting each pulley, as shown in FIG. 6 , each wheel of each pulley according to an embodiment of the present invention includes a wheel flange, a wheel inner surface, a wheel protrusion, a pulley gear, a pressure It may be composed of a sensor film and a slip ring. The wheel flange, the wheel inner surface, the wheel protrusion, the pulley gear, and the pressure sensor film may be configured to rotate integrally with the rotation of the wheel.

The wheel flange means the disk-shaped part with the largest diameter and configured on both sides of the wheel, and is configured perpendicular to the wheel rotation axis, so that the wheel protrusion and the wheel inner part can be configured on the inside, and the pulley gear and slip ring on the outside This can be configured.

The wheel inner portion is configured to be concave in a U shape with a narrower diameter than the wheel flange to connect between the wheel flanges, and the inner wheel portion of the pair of wheels faces each other and may be configured to surround the drawn submarine cable. The wheel protrusion is configured in a direction perpendicular to the outer surface of the inner surface of the wheel to provide friction to the submarine cable, and a pressure sensor film equipped with a pressure sensor may be formed in the center of the inner wheel with the narrowest diameter.

A plurality of wheel protrusions may be configured to protrude in a direction perpendicular to the outer surface of the inner surface of the wheel, and may be configured to extend in a direction horizontal to a rotation axis of the wheel. It can be configured to provide frictional force to the submarine cable by being in direct contact with the sheath of the submarine cable when the cable is pulled out. This can be configured. As shown at the bottom of Figure 6, the wheel protrusion according to an embodiment of the present invention may be configured with a pressure sensor film installation portion concavely configured to be concave as much as the thickness of the pressure sensor film. According to the pressure sensor film installation part according to an embodiment of the present invention, there is an effect that the position of the pressure sensor film does not change despite clamping of the pulley or rotation of the wheel, and all the clamping pressure of the pulley is transmitted to the pressure sensor film This has the effect of reducing the risk of damage to the pressure sensor film.

The pulley gear means a gear configuration that is connected to a gearbox or a chain is connected in order to transmit a driving force transmitted from a driving motor to a specific wheel, and may be configured on the axial outside of one side of the wheel flanges on both sides.

The pressure sensor film refers to a sensor in the form of a film including a pressure sensor circuit and a connection line, and may be coupled to a shape fitting to a pressure sensor film installation part recessed in the wheel protrusion. The connection line of the pressure sensor film is connected to the slip ring on one side of the wheel flange to provide the pressure sensor data to the submarine cable condition monitoring device (1), or is connected to a separately configured communication module and connected to the submarine cable condition monitoring device ( 1) may be configured to provide pressure sensor data.

The slip ring is configured on one side of the wheel flange and is connected to the connection line of the pressure sensor film so that the connection line of the pressure sensor film is not twisted or disconnected despite the rotation of the wheel.

With respect to the modified example of the wheel protrusion and the pressure sensor film, FIG. 7 is a schematic diagram showing a specific configuration of the pulley of the submarine cable pulling-out device according to the modified example of the present invention. 7, in the pulley of the submarine cable pulling device according to a modified example of the present invention, n pressure sensor film installation parts on the wheel protrusion are spaced apart by a specific distance in the direction of the rotation axis to be recessed, and each wheel has n The two pressure sensor films may be configured to be spaced apart by a specific distance in the direction of the rotation axis. According to this, there is an effect that it is possible to more precisely estimate the incoming angle of the submarine cable by the pressure sensor data of the pressure sensor film configured in each wheel.

8 is a schematic diagram showing the operation relationship of the submarine cable condition monitoring device when the submarine cable is pulled out from the offshore wind power generation tower according to an embodiment of the present invention, and FIG. It is a schematic diagram showing the specific configuration of the submarine cable condition monitoring device when the cable is pulled out. As shown in FIG. 8 , the submarine cable condition monitoring apparatus 1 according to an embodiment of the present invention converts driving motor data, pulley rotation data, and pressure sensor data, which are input data for controlling the driving motor, as input data. and subsea cable status data (subsea cable tension overload data, subsea cable angle estimation data, submarine cable tension estimation data, etc.) and optimal control data as output data, Data is transmitted to the operator client through the communication module of the subsea cable condition monitoring device 1, so that the operator can monitor, or it can be configured to directly control the drive motor.

Regarding the specific configuration of the subsea cable condition monitoring device 1, as shown in FIG. 9, the subsea cable condition monitoring device 1 includes a drive motor vector generating module 10, a pulley vector generating module 11, a pressure It may be configured to include a sensor vector generating module 12 , a submarine cable state estimation module 13 , and an optimal control information generating module 14 .

The driving motor vector generation module 10 provides input data (torque, speed, voltage, current, PWM duration, etc.) for controlling the driving motor or measured data outputting the measured state of the driving motor (measured torque, measured speed, etc.) , measured voltage, measured current, etc.) as input data, and driving motor vector as output data.

The pulley vector generation module 11 is an encoding module that receives rotation data of each wheel as input data through an installed RPM sensor, and outputs a pulley vector that is a vector for the entire rotation of each wheel configured in all pulleys as output data. .

The pressure sensor vector generation module 12 receives the pressure sensor data output from the pressure sensor film configured on the wheel protrusion of each wheel as input data, and the pressure that is a vector for the pressure by the submarine cable of each wheel configured in the entire pulley. It is an encoding module that outputs sensor vectors as output data. At this time, the pressure value of each pressure sensor is frequently changed according to the rotation of the wheel. In the pressure sensor vector generating module 12 , the highest pressure value among the pressure values of the pressure sensors constituting each pressure sensor film is set to the corresponding pressure sensor film. It may be configured to pre-process in such a way as to output as a pressure value representative of .

The submarine cable state estimation module 13 is a module that receives a driving motor vector, a pulley vector, and a pressure sensor vector as input data, and outputs the submarine cable state data as output data.

The optimal control information generating module 14 is a module that receives driving motor vector, pulley vector, pressure sensor vector, and submarine cable state data as input data, and outputs optimal control data as output data.

In this case, the driving motor may be composed of a DC motor or an AC motor, and in the case of an AC motor, it may be composed of at least one of an induction motor, a synchronous motor, and a commutator motor, and in the case of a DC motor, a brush motor or a brushless motor (BLDC). It may consist of at least one of When the driving motor is configured as BLDC, it may further include a PID controller, a three-phase inverter for converting DC to three phases, a Hall sensor for sensing the position of the rotor, and commutation logic for determining a switching pattern.

The driving motor data includes input data for controlling the driving motor (torque, speed, voltage, current, PWM duration, etc.) or measurement data outputting the state of the driving motor (measured torque, measured speed, measured voltage, measured current, etc.).

The pulley rotation data may be configured to mean rotation speed or RPM data of each wheel constituting each pulley.

The pressure sensor data may be configured to mean pressure distribution data obtained by merging the pressure values output from the pressure sensor films respectively configured on each wheel of each pulley. At this time, the pressure value of each pressure sensor is frequently changed according to the rotation of the wheel. In the pressure sensor vector generating module 12 , the highest pressure value among the pressure values of the pressure sensors constituting each pressure sensor film is set to the corresponding pressure sensor film. It may be configured to pre-process in such a way as to output as a pressure value representative of . When the highest pressure value among the pressure values of the pressure sensors constituting each pressure sensor film is output as a pressure value representing the corresponding pressure sensor film, the dimension of the pressure sensor vector may be configured to correspond to the total number of wheels. .

The submarine cable state data is the submarine cable tension overload data (for example, a value between 0 and 1), which is the probability value of the class (corresponding to the confidence of the artificial neural network) for whether or not the submarine cable is overloaded with tension, the seabed It may be configured to include submarine cable angle estimation data that is an estimation value of the class for the incoming angle of the cable, and submarine cable tension estimation data that is the estimation value of the class for the tension applied to the submarine cable. According to an embodiment of the present invention, when the submarine cable state data is composed of submarine cable tension overload data, even if the correct submarine cable entry angle or submarine cable tension is not estimated, the worker client is more than a certain confidence level of the submarine cable state data The effect of being able to output a notification to or to limit the output of the driving motor occurs.

The optimal control data is the optimal input data for controlling the driving motor, and is output as specific values of torque, speed, voltage, current, and PWM duration, or the ratio of the optimal input data to the current driving motor input data (for example, For example, it may be composed of 0.9). Furthermore, a plurality of classes are defined as the ratio of the optimal input data to the driving motor input data of the preset current state (eg, ...0.8, 0.9, 1.1, 1.2...), and the optimal control data is each class It can be composed of a probability value (corresponding to the artificial neural network's confidence). According to an embodiment of the present invention, when the optimal control data is composed of the ratio of the optimal input data to the driving motor input data in the current state, it is not necessary to estimate the correct control value, so driving by the subsea cable state monitoring device 1 The effect of facilitating real-time control of the motor occurs.

In relation to a specific embodiment of the submarine cable state estimating module 13, FIG. 10 is a schematic diagram illustrating a case in which the submarine cable state estimating module 13 is configured as a multi-layer neural network. As shown in FIG. 10 , when the submarine cable state estimation module 13 is configured as a multi-layer neural network, the driving motor vector generated by the driving motor vector generating module 10 and the pulley vector generating module 11 An artificial neural network module that uses the generated pulley vector and the pressure sensor vector output from the pressure sensor vector generating module 12 as input data, and an output layer including one output node is configured to use the submarine cable state data as output data. can be configured. In this case, various activation functions such as softmax and ReLU may be used for the output node.

In a training session when the submarine cable state estimation module 13 is configured as a multi-layer neural network, the actual state data (ground truth) and the submarine cable state data output by the submarine cable state estimation module 13 The loss function of the submarine cable state estimating module 13 is configured in a direction to reduce the difference between , and may be configured to update the submarine cable state estimating module 13 .

In an inference session when the submarine cable state estimation module 13 is configured with a multi-layer neural network, a driving motor vector, a pulley vector, and a pressure sensor vector by the driving motor are input as input data, and the current It may be configured to output the submarine cable status data for the status of the submarine cable as output data to transmit the submarine cable status data to the worker client or the optimal control information generating module 14 .

In relation to another embodiment of the submarine cable state estimation module 13, FIG. 11 is a schematic diagram illustrating a case in which the submarine cable state estimation module 13 is configured with an attention module and an RNN encoder. 11, when the submarine cable state estimation module 13 is composed of an attention module and an RNN encoder, the driving motor vector generation module 10 composed of a combination of a CNN encoder and an RNN encoder converts time series driving motor data as input data. It receives the input and outputs the driving motor vector as output data, and the pulley vector generation module 11 composed of a combination of a CNN encoder and an RNN encoder receives the time series pulley rotation data and the driving motor vector as input data and converts the pulley vector as output data. Output, the pressure sensor vector generating module 12 composed of a combination of CNN encoder and RNN encoder is configured to receive time series pressure sensor data and pulley vector as input data and output the pressure sensor vector as output data, the output driving After generating a combined vector by concatenating the motor vector, pulley vector, and pressure sensor vector, the combined vector and the hidden state vector value of the RNN encoder of the submarine cable state estimation module 13 are used in the submarine cable state estimation module 13 ) as the input data of the attention module, and output the attention applied vector as output data, and the outputted attention applied vector and the submarine cable state data of the previous step are composed of the input data of the RNN encoder of the submarine cable state estimation module 13 Thus, it can be configured to output the submarine cable status data of the current step as output data.

According to the combination configuration of the above CNN encoder and RNN encoder of the driving motor vector generating module 10, the pulley vector generating module 11, and the pressure sensor vector generating module 12, the time-series change of the matrix data is outputted as the submarine cable state data. Since it can be used for the subsea cable, there is an effect that the subsea cable condition monitoring device 1 can recognize a sudden change in tension applied to the subsea cable. In addition, when the submarine cable condition estimation module 13 is composed of a combination of the attention module and the RNN encoder as above, attention is applied to the data that is more important for predicting the condition of the submarine cable among the driving motor data, pulley rotation data, and pressure sensor data. Since it is possible to output an attention applied vector according to the situation by giving it, the effect of being able to output the situation-customized submarine cable state data and optimal control data is generated.

12 and 13 are exemplary views of ConvNet (CNN encoder) included in the driving motor vector generating module 10, the pulley vector generating module 11, and the pressure sensor vector generating module 12 according to an embodiment of the present invention. . 12 and 13, taking a simple ConvNet as an example, it can be constructed with [INPUT-CONV-RELU-POOL-FC]. In the case of an input vector, if the INPUT input matrix has 32 horizontal, 32 vertical, and RGB channels, the size of the input may be [32x32x3]. The CONV layer (Conv. Filter) is connected to a partial region of the input matrix, and a dot product of the connected region and its own weight is calculated. The resulting volume will have a size equal to [32x32x12]. The RELU layer is an activation function applied to each element, such as max(0,x). The RELU layer does not change the size of the volume ([32x32x12]). As a result, Activation map 1 is created. The POOL layer (pooling) outputs a reduced volume (Activation map 2) as [16x16x12] by performing downsampling on the "horizontal, vertical" dimensions. The FC (fully-connected) layer calculates the class scores and outputs a pressure distribution vector volume (output layer) with a size of [1x1xn]. The FC layer is connected to all elements of the previous volume.

In this way, the ConvNet included in the driving motor vector generating module 10, the pulley vector generating module 11, and the pressure sensor vector generating module 12 passes through each layer and converts the original matrix composed of pixel values into the class score for the distribution. transform it. Some layers have parameters, some layers have no parameters. In particular, CONV/FC layers are activation functions that include not only input volume but also weight and bias. On the other hand, RELU/POOL layers are fixed functions. The parameters of the CONV/FC layer are learned by gradient descent so that the class score for each matrix is equal to the label of the corresponding matrix.

The parameters of the CONV layer of ConvNet included in the driving motor vector generating module 10, the pulley vector generating module 11, and the pressure sensor vector generating module 12 consist of a series of learnable filters. Each filter is small in the horizontal/vertical dimension, but covers the entire depth in the depth dimension. In the forward pass, each filter is slid (convolve precisely) in the horizontal/vertical dimension of the input volume to generate a two-dimensional activation map. When sliding the filter over the input, a dot product is made between the filter and the input volume. Through this process, ConvNet learns a filter that activates on a specific pattern at a specific location in the input data. Stacking these activation maps in the depth dimension becomes the output volume. Therefore, each element of the output volume treats only a small region of the input, and neurons in the same activation map share the same parameters as they are the result of applying the same filter.

Examples of network structures that can be used in ConvNet included in the driving motor vector generating module 10 , the pulley vector generating module 11 , and the pressure sensor vector generating module 12 are as follows.

LeNet. The first successful ConvNet applications were created by Yann LeCun in the 1990s. Among them, the LeNet architecture that reads zip codes or numbers is the most famous.

AlexNet. What made ConvNet famous in the field of computer vision is AlexNet by Alex Krizhevsky, Ilya Sutskever, and Geoff Hinton. AlexNet participated in the ImageNet ILSVRC challenge 2012 and won 1st place by a big margin (top 5 error rate 16%, 2nd place 26%). The architecture is basically similar to LeNet, but deeper and larger. Also, unlike in the past, where the POOL layer was stacked immediately after one CONV layer, it was configured in a way that several CONV layers were stacked.

ZF Net. The winners of ILSVRC 2013 were Matthew Zeiler and Rob Fergus. It is called ZFNet after the authors. It was created by modifying hyperparameters such as adjusting the size of the intermediate CONV layer in AlexNet.

GoogLeNet. The winner of ILSVRC 2014 was Szegedy et al. This was made by Google. The biggest contribution of this model is to propose an Inception module that greatly reduces the number of parameters (4M, 60M for AlexNet). In addition, at the end of the ConvNet, average pooling is used instead of the FC layer to reduce a lot of seemingly insignificant parameters.

VGGNet. The second-placed network in ILSVRC 2014 is a model called VGGNet created by Karen Simonyan and Andrew Zisserman. The biggest contribution of this model is to show that the depth of the network is a very important factor for good performance. The best of the several models they propose consists of 16 CONV/FC layers, with all convolutions 3x3 and all pooling only 2x2. Although the matrix classification performance is slightly lower than that of GoogLeNet, it was later found that it performs better in several transfer learning tasks. Therefore, VGGNet has been used most recently for matrix feature extraction. The disadvantage of VGGNet is that it uses a lot of memory (140M) and requires a lot of computation.

ResNet. Residual Network created by Kaiming He et al. won the ILSVRC 2015. It uses a peculiar structure called skip connection and has a characteristic that batch normalizatoin is used a lot. This architecture does not use the FC layer in the last layer.

In relation to another specific embodiment of the submarine cable state estimation module 13, FIG. 14 is a schematic diagram illustrating a case in which the submarine cable state estimation module 13 is configured as a multi-layer RNN cell (Recurrent Neural Network cell). 14, when the submarine cable state data is composed of a plurality of data (eg, submarine cable angle estimation data and submarine cable tension estimation data), the RNN encoder of the submarine cable state estimation module 13 is An artificial neural network module that outputs submarine cable angle estimation data and submarine cable tension estimation data according to the layer order, including a plurality of RNN blocks, and using a specific vector as initial input data. The plurality of RNN blocks include a first RNN cell and a second RNN cell, and the first RNN cell receives initial input data or output data of the previous cell, the previous cell RNN hidden layer information, and an attention applied vector to estimate the angle of the submarine cable The data is output, and the second RNN cell may be configured to receive submarine cable angle estimation data and RNN hidden layer information, which are output data of the previous cell, and output submarine cable tension estimation data. 14, the first RNN cell outputs the submarine cable angle estimation data of the current step, and the second RNN cell may be configured to output the submarine cable tension estimation data of the current step. According to the RNN encoder structure of the submarine cable state estimation module 13 according to an embodiment of the present invention, it is configured to sequentially output a plurality of data about the submarine cable state to the layer, There is an effect that the computing resource is reduced.

Regarding a specific embodiment of the optimal control information generating module 14, FIG. 15 is a schematic diagram illustrating a case in which the optimal control information generating module 14 is configured as a reinforcement learning module. As shown in FIG. 14 , the reinforcement learning module included in the optimal control information generating module 14 uses the optimal control information generating module 14 as an Agent, driving motor data of the current step, pulley rotation data, pressure The sensor data is the State, the submarine cable status data or attention applied data is the Environment, and the optimal control data generated by the optimal control information generating module 14, which is the agent, in the state in the environment, is the action. And, as the ratio value of the drive motor data and the actual cable pulling-out speed or the ratio value of the pulley rotation data and the actual cable pulling-out speed is relatively large, a higher reward is generated to update the optimal control information generating module 14 that is the agent. can In the learning session of the optimal control information generating module 14, the actual cable pulling-out speed may be configured to be applied to the submarine cable using a marker or an acceleration sensor, and in the inference session of the optimal control information generating module 14, learning is completed. The hidden layer parameter of the optimal control information generating module 14 may be configured to be fixed. When the optimal control information generating module 14 is configured as a reinforcement learning module, it is possible to quickly output optimal control data without using an optimal algorithm with high complexity, so that the driving motor is controlled in real time according to the tension load of the submarine cable. There is an effect that can be done.

As described above, those skilled in the art to which the present invention pertains will be able to understand that the present invention can be implemented in other specific forms without changing the technical spirit or essential features thereof. Therefore, the above-described embodiments are to be understood in all respects as illustrative and not restrictive. The scope of the present invention is indicated by the following claims rather than the detailed description, and all changes or modifications derived from the meaning and scope of the claims and equivalent concepts should be construed as being included in the scope of the present invention.

The features and advantages described herein are not all inclusive, and many additional features and advantages will become apparent to those skilled in the art, particularly upon consideration of the drawings, the specification, and the claims. Moreover, it should be noted that the language used herein has been principally selected for readability and teaching purposes, and may not be chosen to delineate or limit the subject matter of the present invention.

The foregoing description of embodiments of the present invention has been presented for purposes of illustration. It is not intended to be exhaustive or to limit the invention to the precise form disclosed. Those skilled in the art will appreciate that many modifications and variations are possible in light of the above disclosure.

Therefore, the scope of the present invention is not limited by the detailed description, but by any claims of the application based thereon. Accordingly, the disclosure of the embodiments of the present invention is illustrative and not intended to limit the scope of the present invention as set forth in the following claims.

1: Submarine cable condition monitoring device
10: drive motor vector generation module
11: Pulley Vector Generation Module
12: pressure sensor vector generation module
13: submarine cable condition estimation module
14: Optimal control information generation module

Claims (10)

In the device for monitoring the status of the submarine cable in the submarine cable pulling device for pulling the submarine cable in order to lead the submarine cable in from the offshore wind power tower,
The submarine cable pulling device includes a plurality of pulleys composed of two pairs of wheels facing each other and a driving motor for providing a driving force to the wheels of the pulleys, the foundation structure of the offshore wind power tower It is configured so that the submarine cable insertion part of the pulley is located in the submarine cable entry part of the offshore structure configured above,
The driving motor input data including torque, speed, voltage, current or PWM duration for controlling the driving motor or the measured state of the driving motor including the measured torque, the measured speed, the measured voltage or the measured current a driving motor vector generating module, which is an encoding module that uses driving motor data including output driving motor measurement data as input data, and encodes the driving motor data to use the driving motor vector as output data;
Using pulley rotation data, which is rotation data of each wheel constituting the pulley, as input data, encoding the pulley rotation data, and outputting a pulley vector that is a vector for rotation of each wheel configured in the pulley as output data an encoding module, a pulley vector generation module; and
a submarine cable state estimation module using the driving motor vector and the pulley vector as input data and using the submarine cable state data, which is data on the state of the submarine cable, as output data;
including,
It is connected to a worker client or a controller of the drive motor through a wired/wireless network, and transmits the submarine cable status data to the worker client or controls the controller of the drive motor based on the submarine cable status data,
A subsea cable condition monitoring device when a subsea cable is pulled out from an offshore wind power tower.
According to claim 1,
The submarine cable pulling device further includes a pressure sensor film configured inside the wheel to output pressure sensor data,
A pressure sensor vector generating module that uses the pressure sensor data as input data, encodes the pressure sensor data, and uses the pressure sensor vector as output data;
The submarine cable state estimation module, characterized in that the pressure sensor vector, the driving motor vector and the pulley vector as input data,
A subsea cable condition monitoring device when a subsea cable is pulled out from an offshore wind power tower.
3. The method of claim 2,
The submarine cable drawing device further includes a pressure sensor film installation part configured to be recessed to fit the pressure sensor film on the inside of the wheel,
The pressure sensor film is characterized in that it is custom-coupled to the pressure sensor film installation part,
A subsea cable condition monitoring device when a subsea cable is pulled out from an offshore wind power tower.
4. The method of claim 3,
In the submarine cable pulling device, a plurality of the pressure sensor film installation parts are spaced apart by a specific distance in the direction of the rotation axis of the pulley,
The pressure sensor film, characterized in that it is configured in plurality and is custom-coupled to each of the plurality of pressure sensor film installation parts,
A subsea cable condition monitoring device when a subsea cable is pulled out from an offshore wind power tower.
According to claim 1,
The submarine cable state data, characterized in that it includes data on whether the tension of the submarine cable is overloaded, which is a probability value for whether the tension of the submarine cable is greater than or equal to a specific value,
A subsea cable condition monitoring device when a subsea cable is pulled out from an offshore wind power tower.
According to claim 1,
an optimal control information generating module using the driving motor vector and the pulley vector as input data and using optimal control data, which is data for control of the driving motor, as output data;
It is connected to the controller of the driving motor through a wired/wireless network, characterized in that it controls the controller of the driving motor based on the optimal control data,
A subsea cable condition monitoring device when a subsea cable is pulled out from an offshore wind power tower.
According to claim 1,
The submarine cable state estimating module is composed of a multi-layer neural network using the driving motor vector and the pulley vector as input data, and using the submarine cable state data, which is data on the state of the submarine cable, as output data, one An output layer including an output node of the is configured to make the submarine cable state data output data,
In a training session of the submarine cable state estimation module, the submarine cable state estimation module in the direction of reducing the difference between the actual state data (ground truth) and the submarine cable state data output by the submarine cable state estimation module characterized in that the loss function of is configured so that the submarine cable state estimation module is updated,
A subsea cable condition monitoring device when a subsea cable is pulled out from an offshore wind power tower.
According to claim 1,
The submarine cable state data includes submarine cable angle estimation data and submarine cable tension estimation data,
The submarine cable state estimation module includes a first RNN cell and a second RNN cell,
The first RNN cell receives the output data of the previous cell, the RNN hidden layer information of the previous cell, and the input data of the submarine cable state estimation module and outputs the submarine cable angle estimation data, and the second RNN cell is the previous cell characterized in that it is configured to output the submarine cable tension estimation data by receiving the submarine cable angle estimation data and the RNN hidden layer information of the first RNN cell as output data of
A subsea cable condition monitoring device when a subsea cable is pulled out from an offshore wind power tower.
According to claim 1,
an optimal control information generating module using the driving motor vector and the pulley vector as input data and using optimal control data, which is data for control of the driving motor, as output data;
The optimum control information generation module uses the optimum control information generation module as Agent, the driving motor data of the current step and the pulley rotation data as State, and the submarine cable state data as Environment, and states in the Environment. In the above, the optimal control data output by the optimal control information generating module, which is the Agent, to the controller of the driving motor is taken as an action, and the ratio value of the driving motor data and the actual cable pulling-out speed or the pulley rotation data and the actual cable A higher reward is generated as the ratio value of the drawing speed is relatively large, and the optimal control information generating module, which is the agent, is configured to be updated,
It is connected to the controller of the driving motor through a wired/wireless network, characterized in that it controls the controller of the driving motor based on the optimal control data,
A subsea cable condition monitoring device when a subsea cable is pulled out from an offshore wind power tower.
In the method of monitoring the status of the submarine cable in the submarine cable pulling device for pulling the submarine cable to lead the submarine cable in from an offshore wind power tower,
The submarine cable drawing device includes a plurality of pulleys composed of two pairs of wheels facing each other and a driving motor for providing a driving force to the wheels of the pulleys, the foundation structure of the offshore wind power tower It is configured so that the submarine cable insertion part of the pulley is located in the submarine cable entry part of the offshore structure configured above,
The driving motor vector generation module includes driving motor input data or measured torque, measured speed, measured voltage or measured current including torque, speed, voltage, current or PWM duration for controlling the driving motor a driving motor vector generating step of using as input data driving motor data including driving motor measurement data for outputting the measured driving motor state, encoding the driving motor data and outputting the driving motor vector as output data;
The pulley vector generation module takes as input data pulley rotation data, which is rotation data of each of the wheels constituting the pulley, as input data, encodes the pulley rotation data, and is a pulley vector that is a vector for rotation of each wheel configured in the pulley. a pulley vector generation step of outputting as output data; and
a submarine cable state estimating step in which the submarine cable state estimating module uses the driving motor vector and the pulley vector as input data, and outputs the submarine cable state data, which is data on the state of the submarine cable, as output data;
including,
It is connected to a worker client or a controller of the drive motor through a wired/wireless network, and transmits the submarine cable status data to the worker client or controls the controller of the drive motor based on the submarine cable status data,
A method of monitoring the condition of a submarine cable when it is pulled out from an offshore wind power tower.

Images