Method and system for improving reliability of dredger sensing operation system

Abstract

To address the problem of construction discontinuities caused by critical parameter sensor failures in the harsh construction environment of cutter suction dredgers, this paper proposes a method and system for improving the reliability of the dredger's sensing operation system. This method establishes a sensor network regression prediction model based on the "digital twin" concept. This model uses a stacking algorithm to establish a correlation model between the dredger's critical parameter sensors and other high-reliability sensors. This model can synchronously calculate the "twin" data of any critical parameter sensor during dredger operation, serving as a backup for the key parameter sensors. The "Long Lion 12" CSD construction case demonstrates that this method has high prediction accuracy, with all key parameter predictions exceeding 0.9 in goodness-of-fit, effectively addressing the problem of construction discontinuities caused by critical sensor failures.

Description

Technical Field

[0001] The present invention relates to the application of new generation information technology on a cutter suction dredger, and in particular to a method and system for improving the reliability of a sensing operation system of a dredger. Background Art

[0002] Cutter suction dredgers, a type of dredging vessel, offer advantages such as high efficiency, low cost, adaptability to diverse soil types, flexible operation, strong maneuverability, suitable dredging depth, and environmentally friendly dredging operations. They are widely used in land reclamation, waterway maintenance, and port desilting. Digitalizing dredging operations can improve dredging efficiency, reduce construction costs, and significantly reduce the workload of construction personnel. Digitalizing dredging operations requires efficient and reliable monitoring and sensing equipment and timely and stable automatic control systems. However, dredging sites are subject to harsh environments, and sensors operating in high temperature and high humidity conditions year-round are prone to failure. This can lead to data acquisition failures or transmission interruptions, impacting the continuity of dredger operations. Therefore, the stable and effective collection of key dredger operating parameters has long been a research hotspot in the dredging field.

[0003] During the operation of a cutter suction dredger, operators primarily rely on monitoring key parameters such as mud concentration, mud flow rate, underwater pump vacuum, mud pump discharge pressure, cutter power, underwater pump discharge pressure, and traverse speed to determine the operation status. In practice, due to the harsh environment of high temperature, high humidity, and high corrosion, the dredger's underwater pump suction vacuum sensor, underwater pump discharge pressure sensor, mud pump discharge pressure sensor, flow sensor, and concentration meter often malfunction, affecting the operator's judgment of the cutter suction dredger's operation status and, in severe cases, causing construction to be interrupted. Summary of the Invention

[0004] This invention provides a method and system for automatically selecting construction sites for a cutter suction dredger based on global optimization. This method can help a dredger autonomously complete its dredging task within a specified timeframe without requiring human intervention. Regardless of any unforeseen circumstances, this method ensures that the dredger completes its construction task on schedule.

[0005] According to a first aspect of an embodiment of the present invention, a method for improving the reliability of a dredger's sensing operation system is provided, comprising: calculating the correlation between the dredger's sensors and using this correlation to establish a networked model between a target sensor and other sensors; and utilizing this networked model to promptly correlate the target sensor's data under current operating conditions when the target sensor fails. When the target sensor is operating normally, the networked model also outputs twin data along with the target sensor, forming a primary-auxiliary backup. The maximum mutual information coefficient method is used to calculate the correlation between the sensors and select the most relevant sensor corresponding to each sensor.

[0006] According to a second aspect of an embodiment of the present invention, a system for improving the reliability of a dredging vessel's sensing operation system is provided, comprising: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to execute all or part of the steps of the method.

[0007] According to a third aspect of an embodiment of the present invention, a non-transitory computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the computer program implements all or part of the steps of the method. The non-transitory computer-readable storage medium may be a ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage device, or the like. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings of the embodiments are briefly introduced below.

[0009] Figure 1 A conceptual diagram of a twin sensor provided in accordance with one embodiment of the present invention.

[0010] Figure 2 A diagram of the maximum mutual information coefficient calculation process provided by one embodiment of the present invention.

[0011] Figure 3 An embodiment of the present invention provides a diagram for calculating the maximum cross-correlation coefficient of each sensor.

[0012] Figure 4 A flowchart of a Stacking algorithm provided in one embodiment of the present invention.

[0013] Figure 5 A schematic diagram of a twin sensor network for a dredger provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0014] To address the problem of construction discontinuity caused by critical parameter sensor failures in the harsh construction environment of cutter suction dredgers, this paper proposes a sensor network regression prediction model based on the concept of "digital twins." This model uses a stacking algorithm to establish a correlation model between the dredger's critical parameter sensors and other high-reliability sensors. This model can synchronously calculate the "twin" data of any critical parameter sensor while the dredger is operating, serving as a backup for the key parameter sensors. The "Long Lion 12" CSD construction case verified that this method has high prediction accuracy, with all key parameter predictions having goodness-of-fit values ​​exceeding 0.9, effectively resolving the construction discontinuity caused by critical sensor failures.

[0015] The networking method for stable output of sensor data based on digital twin technology proposed in this invention uses big data to analyze the intrinsic relationship between different sensors and establish a networking model between the target sensor and other high-reliability sensors. That is, when the target sensor fails, the network model can be used to timely associate the target sensor data under the current working conditions. Its concept is as follows Figure 1 As shown, when Figure 1 If sensor A fails, the established sensor network can quickly replace it with its "twin sensor" A', maintaining smooth sensor network operation. When the target sensor is functioning properly, the sensor network also outputs a "twin" data, forming a primary-auxiliary backup to verify the reliability of the actual sensor. This sensor networking technology can cleverly address issues that affect dredger operation continuity caused by sensor failures, significantly improving dredger operation efficiency.

[0016] For the purpose of learning all the historical construction data of the dredger, the data is first pre-processed in advance (including noise removal, normalization, etc.), and then the maximum mutual information coefficient method is used to calculate the correlation between the sensors, and the most relevant sensor category corresponding to each sensor is selected. The specific calculation process is as follows: first, assume that the set of data points with two attributes is distributed in a two-dimensional space, and use an m by n grid to divide the data space, so that the frequency of the data points falling in the (x, y) grid is used as an estimate of P(x, y), and the frequency of the data points falling in the xth row is used as an estimate of P(x), and similarly, an estimate of P(y) is obtained. Then calculate the mutual information of random variables X and Y. Because there is more than one way to divide the data points into an m by n grid, this application aims to obtain the grid division that maximizes the mutual information. Then use the normalization factor to convert the mutual information value into the (0, 1) interval. Finally, find the grid resolution that maximizes the normalized mutual information as the measure of the maximum mutual information coefficient.

[0017] Therefore, the calculation of the maximum mutual information coefficient is divided into three steps, such as Figure 2 As shown:( Figure 2 A) Given i and j, grid the scatter plot composed of X and Y and find the maximum mutual information value; ( Figure 2 B) Divide the maximum mutual information value by log2(min(X|,|Y)) to obtain a normalized value; ( Figure 2 C) Find the maximum mutual information value at different scales as the maximum mutual information coefficient value.

[0018] Mathematical expression:

[0019]

[0020]

[0021] Figure 3 The figure shows the correlation between the two parameters. The white part in the figure indicates that there is no correlation between the two parameters, and both are constant values.

[0022] After each sensor selects its corresponding related sensor, it is input into the Stacking algorithm model for training and learning. The Stacking algorithm is a hierarchical integrated learning framework that uses a two-layer model. The first layer consists of multiple base learners, and the input is the original training set. The meta-learner in the second layer is retrained with the output of the first layer base learner as the training set to obtain a complete Stacking model. The specific pseudo code is shown in Table 1, and the process is as follows Figure 4 .

[0023] After the training is completed, the network composed of sensors of the dredger is as follows Figure 5 As shown in the figure, the sensor network corresponding to the F20 mud flow and the F13 reamer speed is shown. It can be clearly seen in the figure that if F20 or F13 fails, the theoretical value of the failed sensor can be quickly calculated using other sensors.

[0024] Table 1

[0025] Stacking pseudocode

[0026]

[0027]

Claims

1. A method for improving the reliability of a dredger's sensing operation system, characterized in that: include: The maximum mutual information coefficient method is used to determine the correlation between sensors: the data points in the two-dimensional space are divided into an m×n grid, and the mutual information is calculated by statistically analyzing the frequency distribution of P(x, y), P(x), and P(y). The mutual information value is mapped to the (0, 1) interval using a normalization factor. selecting a grid resolution that maximizes normalized mutual information as a maximum mutual information coefficient; and selecting the most relevant sensor category for each sensor based on the maximum mutual information coefficient; Based on the selected relevant sensor categories, a stacking model was constructed using a hierarchical ensemble learning framework. The first layer consists of multiple base learners, whose input is preprocessed historical construction data containing correlations between sensors. The second layer meta-learner is retrained using the output of the first layer base learners as the training set. The trained Stacking model is used as the core architecture of the sensor networking model to achieve dynamic data association between the target sensor and related sensors; When a target sensor fails, the target sensor data under the current working condition is correlated in real time through the Stacking model; When the target sensor operates normally, the Stacking model outputs twin backup data synchronized with the target sensor data, forming a primary-secondary backup mechanism.

2. A system for improving the reliability of a dredger's sensing operation system, characterized in that: include: processor; a memory for storing processor-executable instructions; The processor is configured to execute the steps of the method according to claim 1.

3. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to claim 1 are implemented.

Images