Application of improved genetic algorithm in barge loading of offshore platform

Abstract

In order to complete the loading operation in a short time, the genetic algorithm was improved, and the improved mechanisms such as the total crossover of the population and the distributed dynamic penalty function method were proposed. The distribution of ballast water in barge stowage was optimized. The calculation of the ship’s floating state, stability, the check of the program strength, and the improvement of the genetic algorithm were discussed. By simplifying the model, all ballast water tanks were involved in the stowage, so as to better meet the limitation of tide level, tank capacity and strength. A more adaptable load optimization solution was obtained. The results show that the improved genetic algorithm has shorter time and higher efficiency than the basic loading scheme. The improved genetic algorithm was used to optimize the barge loading plan, which met the engineering requirements and shorten the working time of the barge. Therefore, the research on the improved genetic algorithm can effectively improve the quality and reliability of the project, which is beneficial to improve the scientific, safety and reliability of the barge loading operation. This program has important implications for marine engineering.

Keywords

Genetic algorithms offshore platform barge stowage

1 Introduction

Barge loading is an important part of the construction process of large marine structures. There are many factors to consider in this technology. It is a difficult job. All countries in the world have paid attention to the important technical exploration of this operation [1]. With the vigorous development of the offshore oil industry, the development of the sea area has moved from the shallow sea to the deep sea, and the offshore platform is developing in a higher direction. The transfer of these large-scale offshore structures for deep-sea operations is highly demanding on barge loading technology. Usually, marine structures are loaded by floating cranes. However, this method is only suitable for small and medium-sized structures. For large structures, lifting operations cannot be carried out [2]. In order to solve the problem of loading and unloading of large-scale marine structures, the barge loading technology for this situation has been specifically proposed in the project. With the development and application of computer technology in various industries, a ship-loading system based on computer technology has been produced [21 –28]. In general, the technology has achieved initial results in engineering operations, but its technological development needs to be improved. There are still some areas for improvement [3]. For example, when the working time is limited, a reliable and practical loading algorithm is needed. The barge stowage operation mainly uses the information of the barge products, the barge itself and the surrounding environment. The basic principles of ship statics and ship strength are used for calculation, including the calculation of ship’s buoyancy, stability and strength [4].

At present, there are not many researches on barge loading at home and abroad. The general work is only to meet the requirements of the current situation of the project. Moreover, in the complex situation of the site, faced with changing factors, the solution in the project may sometimes face a situation without a solution, which has to rely on the staff’s experience again. Therefore, it is particularly critical to write a good barge stowage algorithm, and it is very important to find a reasonable scheme that meets the engineering requirements among various schemes. The improved genetic algorithm was used to optimize the barge stowage scheme, and the best scheme of each stowage was found among many feasible schemes, so as to meet the engineering requirements reasonably.

2 Literature review

Many scholars’ literature on ship stowage mainly focuses on container ship stowage planning. In 2015, Ding and Chou proposed a heuristic algorithm for container ship stowage planning problems. This algorithm abstractly considers the problem of loading planning for container ships. Container ships access a series of ports in sequence, and containers can only enter from the top of the container [5]. In 2015, Morris et al. explored the application of neural network expert systems in cargo ship loading systems by combining artificial neural networks with expert systems [7]. In 2015, Elhoseny et al. studied the inefficiency of manual loading of Baosteel’s water transport business. By analyzing the ship loading principle and combining the ship’s requirements for strength, stability and load, a ship loading expert system has been developed and achieved good results [8]. In 2015, Niesse and Mayne developed a ship loading fuzzy optimization expert system using fuzzy mathematics theory and expert system theory. Accurate, fast and reasonable loading of the goods is realized, and the loading map can be quickly drawn [9]. In 2015, Rauwolf used the real-time transfer of water storage in the ship’s transfer tank to control the lateral tilt caused by the change of the ship’s loading during the voyage, thus ensuring the safety of the ship [14]. In terms of airbag transport, in 2015, Hai et al. studied the application of airbag haulage technology in port engineering [15]. In 2015, Kai et al. introduced the technology of airbag transporting caisson, which provided a technical reference for the construction of stowage [16]. In 2015, Ko et al. used engineering mechanics theory to quantify the water transfer of the floating dock during the shipping process [17].

The quantitative calculation of the lap joint force during the delivery process reduces the risk of qualitative analysis based on experience in the past. In 2015, Xin et al. proposed and developed the automatic loading procedure of the ship, and gave the calculation results of the ship stability, floating state, strength and other indicators that meet the relevant specifications and conventions [18]. In 2015, Goerlandt and Montewka effectively solved the problem of residual strength reduction of oil tankers due to aging by rationally controlling the tanker loading scheme and loading sequence [19]. In 2016, Chaves and Leduino proposed that the container ship loading plan problem is an important issue in the operation of port container terminals [6]. In 2016, Chin et al. regarded the preparation of optimal ballast scheme for ships as a multi-objective decision-making problem. The basic principle of the optimal ballast scheme is given, which improves the user’s operational efficiency [13]. In 2017, based on the research of container ship full-line loading, using intelligent optimization algorithms and techniques, some new methods for solving the loading problem were proposed [10]. In 2017, Moan and ayala-uraga carried out quantitative calculation of the optimal trim, ballast and stowage of the ship based on the principle of optimal trim navigation energy saving. It has been proved that the fuel economy rate of this method can reach 2% -3% [12]. In 2018, Feng et al. proposed new ideas and methods from the perspective of improving equipment efficiency. That is, based on the pre-loading map limitation, an integer linear programming model is established. The container group on the yard is most reasonably distributed to the ship’s unit block [11]. In 2018, Ibrahim discussed the method of computer-aided loading. This stowage system can be used to calculate and check the buoyancy, stability and strength of manual tanker stowage scheme [20].

3 Methodology

Genetic algorithm is a global optimization probability search algorithm that simulates the behavior of biological evolution through genetics and mutation in nature. It was first proposed by Professor Holland of the University of Michigan, USA. On the basis of summarizing the work of predecessors, the basic framework of genetic algorithm is proposed. At the beginning of the genetic algorithm, an initial population needs to be generated. Then, according to the search rules of the algorithm, the new population is obtained by selecting, crossing and mutating the population. At this point, a generation of evolution is complete. The evolutionary population is judged by the stopping mechanism. If the stopping requirement is met, the evolution process is exited, the evolution result is output, and the calculation process is completed. The genetic algorithm can be expressed as an eight-element model, as shown in Equation (1): $SGA = (code, fitness, pop 0, popsize, sel, cross, muta, T)$ (1)

In the formula, code represents the coding mode adopted by the algorithm, and fitness represents the fitness evaluation function of the individual. pop0 represents the initial population of the algorithm, and popsize represents the number of individuals in the population. sel represents the selection operator and corss represents the crossover operator. muta represents the mutation operator. T represents the termination condition of the algorithm. It is often taken as the largest evolutionary algebra of the population. The execution flow chart of the model is shown in Fig. 1:

Fig. 1

Basic process of genetic algorithm.

To study the loading problem of the launching barge, the loading plan optimization model is the constrained optimization model. The constraint conditions are the equation constraints of equilibrium equations and the upper and lower limits of regulating water volume. Using the method of distributed dynamic parameter penalty function to deal with the constraint, the fitness value of the population is calculated as follows: $fitness = f (X) • λ • β • Rf (X)$ (2)

In the formula, f(X) is the individual objective function value, and Rf(X) is the distance that the individual exceeds the constraint.λ is the adjustment factor, which is used to adjust the primary and secondary problems between the optimization target and the constraint. It is a fixed value. β is the scale factor of the segmentation penalty, which is determined by the degree of individual deviation from the constraint and the distance allowed to exceed the constraint. According to the multiple relationship of the distances deviating from the constraint, β is taken as seven values between 0.8 and 2, respectively.

4 Results and discussion

4.1 Floating, stability and strength of ships

In the calculation of the ship’s floating state, draught is the first problem to be solved. In general, the head and tail draft are not equal, which formed the draft difference. The draught difference can be determined by the centimeter pitching moment Mcm and the ship trim moment ML in the hydrodynamic curve. That is, the difference between the first and last drafts of the ship’s trim can be expressed as: $t = \frac{D (x_{g} • x_{b})}{100 M_{cm}}$ (3)

Among them, xg is the ordinate of the center of gravity of the ship, and xb is the coordinate value of the ship’s floating center along the length of the ship. According to the ship coordinate system, the front of the ship is positive, and the middle of the ship is negative. In the calculation of the ship’s floating state, it is important to determine the draught, displacement and position of the ship’s buoyancy. The primary problem is to calculate the cross-sectional area of the ship and its static moment. Usually, Bonjean’s curves query method is used. Here, the Green’s formula method is used. In engineering calculations, there are often many repeated calculations for the elements below the ship’s waterline. A lot of energy is wasted on these trivial and complicated jobs. In order to facilitate staff’s calculation and reference, some of these elements are generally drawn into multiple curves, which are called Bonjean’s curves. The ship’s buoyancy curve along the length of the ship under a certain load condition is called the ship’s buoyancy curve. The ship’s buoyancy curve can express the following information: The area under the buoyancy curve is equal to the buoyancy acting on the hull. The area centroid longitudinal coordinate value of the buoyancy curve is the position of the floating center in the longitudinal direction of the ship. In general, the buoyancy curve distribution of a ship is not in a state of horizontal buoyancy, so the ship has to adjust its trim after the actual balance position is obtained. In barge stowage operation, the state of the barge is known at a specific moment, and the final effect to be achieved is the flat float state. Therefore, no trim adjustment is required. After obtaining the draught of the barge, the buoyancy curve of the barge corresponding to the floating state can be drawn by querying the cross-sectional area of each station under the load condition through the state curve.

4.2 Barge information and data processing

Since the loading model is a large program, the amount of data exchange is relatively large. The loading principle and the calculation effect of the algorithm are explained only in one loading situation of the barge. Taking a large barge as an example, the process of its stowage is analyzed.

The barge products are generally large marine structures. The loading and unloading modes can be generally divided into airbag transportation and slide transportation. Here, only the slipway transfer is considered. When the barge chute is aligned with the pier chute, the large structure advances with the continuous winding of the winch. As the structure advances, the points and forces acting on the barge are constantly changing. In order to simplify the problem without changing the nature of the barge structure’s force on the barge, the weight of the structure is distributed in a segmented rectangle, as shown in Fig. 2. Among them, the product’s coordinate system is shown in the figure. The intersection line between the middle longitudinal section and the bottom plane is the x-axis, and the direction is positive along the propulsion direction. The intersection of the middle cross section and the ground plane is the y-axis, and the starboard side is positive. The intersection of the middle longitudinal section and the middle cross section is the z-axis, and the direction is vertical and upward. In the Fig. 1, the gravity load at point xi is gi, the weight of the barge product is Gp, and the center of gravity is O’.

Fig. 2

Distribution of weight of barter products.

Fig. 3

Concise layout of barge ballast bay.

In the calculation of barge loading, draft is that the floating state of barge reaches the predetermined standard. At the same time, the strength of the barge is calculated. It is not allowed to exceed the shear moment limit in the process of water diversion. If there is such a scheme, it must be abandoned. Then, other schemes are calculated and their shear moment values are verified until a reasonable solution is found. As shown in Table 1.

Table 1

Barring product parameters

Product Elements
Weight (t)	Center of gravity(x)	Center of gravity(y)	Center of gravity(z)
7000	–10.238	0.0	4.5
Long(m)	Wide(m)	Type depth(m)	Product entry length(cm)
73.22	55.78	7.62	8.0

4.3 Brief introduction of basic loading method

The final goal of barge stowage is to adjust the amount of water. The result is the difference between the next preset moment and the current barge buoyancy state, including the force and torque difference generated by adjusting the amount of water. The prototype of the basic barge loading method is the pairing method proposed in the doctoral thesis published by Sun Chengmeng. Using the special structure of the barge, the simplification of the adjustment of the core of the water tank is carried out in the treatment of the lateral and longitudinal loading of the ballast tank. Specifically, the two rows of water tanks to be adjusted are regarded as a whole when the barge is vertically adjusted. Among them, the adjusted water volume of a row of cabins in each loading scheme is used as a design variable. According to the knowledge of statics, the resultant force and resultant moment of the barge after the scheme stowage are zero. Therefore, this problem is transformed into a solution problem of binary quadratic equations. The method of barge loading lateral adjustment is the same. The adjustment time for each step of the designed barge stowage scheme depends on the maximum regulated water volume in the two train compartments.

4.4 Comparison between electric heating and fire exposure

The maximum adjusted water volume of the improved genetic algorithm loading scheme is smaller than the maximum adjusted water volume of the basic loading scheme. The improved genetic algorithm has less adjusting time for the corresponding stowage scheme. It can be seen that under the same parameters, the improved barge loading optimal solution calculated by the genetic algorithm can save the time of the loading process. In this barge stowage, the error of regulating water quantity, the trim moment error and the sum of the corresponding heeling moment error are less than the given threshold. From the analysis of the calculated data, the error is within the required range. Each optimization scheme is compared again, and the scheme with the smallest amount of regulating water in all optimization schemes is selected as the preliminary scheme. The preliminary plan is to improve the corresponding scheme of the genetic algorithm in the first scheme. The maximum shear moment generated by this solution is calculated. The program meets all requirements. Therefore, this solution is the optimal solution for this stowage, that is, the final plan.

5 Conclusion

The barge stowage problem can be essentially treated as a constrained optimization problem. From the engineering process, in general, the convenience of pump control will be given priority. A column of cabins is used as the basic selection unit. Several cabins in a limited column (typically two columns) are selected to regulate the water tank. The problem is simplified (with little effect on the result) into an equilibrium equation that can be solved accurately to obtain the barge stowage scheme. The improved genetic algorithm is used to optimize the barge loading, which can meet the engineering requirements and shorten the working time of the barge. In the optimization process of this algorithm, multiple subgroup parallel computing modes in the niche genetic algorithm are applied. This has a good effect on the multi-peak case processing of the function. After calculation and verification, the improved genetic algorithm has a good effect. As shown in Table 2.

Table 2

Current barge cabin depth

	Barge ballast parameters
	Column1	Column2	Column3	Column4
	(m)	(m)	(m)	(m)
Starboard	4.896	2.532	2.532	2.532
	2.279	2.532	2.532	2.532
Port	2.245	2.532	2.532	2.532
	4.896	2.532	2.532	2.532
	Column 5	Column 6	Column 7	Column 8
	(m)	(m)	(m)	(m)
Starboard	2.532	1.0	4.422	5.5
	2.532	1.0	2.358	5.5
Port	2.532	1.0	2.355	5.5
	2.532	1.0	4.419	5.5

From the point of view of genetic algorithm, the loading scheme of barge is optimized. Although the calculation results obtained are ideal, the algorithm also has some shortcomings. The optimization algorithm is a linear programming problem. When considering the ballast tank core, simplified treatment is adopted. The core of ballast tank before stowage is directly used to replace the core of ballast tank after stowage. The calculation result has a slight error. If the curves of each tank capacity can be used to simulate the curve of smoothness, and the shape center after load adjustment can be used to participate in the calculation, the calculation results will be more reliable.

References

Bharathi

, Rekha

and Vijayakumar

, Genetic algorithm based demand side management for smart grid, Wireless Personal Communications 93(2) (2017), 481–502.

Norman

B.A.

and Bean

J.C.

, A genetic algorithm methodology for complex scheduling problems, Naval Research Logistics 46(2) (2015), 199–211.

Hartmann

, A competitive genetic algorithm for resource-constrained project scheduling, Naval Research Logistics 45(7) (2015), 733–750.

Aickelin

and Dowsland

K.A.

, Exploiting problem structure in a genetic algorithm approach to a nurse rostering problem, Journal of Scheduling 3(3) (2015), 139–153.

Ding

and Chou

M.C.

, Stowage planning for container ships: A heuristic algorithm to reduce the number of shifts, European Journal of Operational Research 246(1) (2015), 242–249.

Chaves

A.A.

, Leduino

D.S.N.L.

, et al., Pareto clustering search applied for 3D container ship loading plan problem, Expert Systems with Applications 4 (2016), 50–57.

, Nonlinear information data mining based on time series for fractional differential operators, Chaos 29 (2019), 013114. DOI: 10.1063/1.5085430

, Wu

, Wang

and Zou

, Evaluation of developer efficiency based on improved DEA model, Wireless Personal Communications 102(4) (2018), 3843–3849.

, Wang

and Yuntao

, Sewage information monitoring system based on wireless sensor, Desalination and Water Treatment 121 (2018), 73–83.

10.

Tavakkolimoghaddam

, Safari

and Sassani

, Reliability optimization of series-parallel systems with a choice of redundancy strategies using a genetic algorithm, Reliability Engineering & System Safety 93(4) (2017), 550–556.

11.

Feng

, Butler-Purry

K.L.

and Zourntos

, Real-time electric load management for DC zonal all-electric ship power systems, Electric Power Systems Research 154 (2018), 503–514.

12.

Moan

and Ayala-Uraga

, Reliability-based assessment of deteriorating ship structures operating in multiple sea loading climates, Reliability Engineering & System Safety 93(3) (2017), 433–446.

13.

Chin

H.M.

, Su

C.L.

and Liao

C.H.

, Estimating power pump loads and sizing generators for ship electrical load analysis, IEEE Transactions on Industry Applications 52(6) (2016), 4619–4627.

14.

Rauwolf

G.A.

, Near-optimal low-thrust orbit transfers generated by a genetic algorithm, Journal of Spacecraft & Rockets 33(6) (2015), 859–862.

15.

Hai

, Wen

, Hong

Y.Y.

, et al., Optimal sizing of hybrid PV/diesel/battery in ship power system, Applied Energy 158 (2015), 26–34.

16.

Kai

, Sun

and Yi

, Research on adaptive guidance technology of UAV ship landing system based on net recovery, Procedia Engineering 99 (2015), 1027–1034.

17.

D.E.

, Shin

S.H.

and Cho

Y.J.

, A study on the application of new lining material for ship propulsion shafting system, IOP Conference Series: Materials Science and Engineering 269 (2017), 12–13.

18.

Xin

, Rui

, Wang

, et al., Vibration control of shipborn multiple launch rocket systems, Journal of Nanjing University of Science & Technology 39(1) (2015), 1–6.

19.

Goerlandt

and Montewka

, A framework for risk analysis of maritime transportation systems, A case study for oil spill from tankers in a ship–ship collision, Safety Science 76 (2015), 42–66.

20.

Ibrahim

R.A.

, Overview of structural life assessment and reliability, part III: Impact, grounding, and reliability of ships under extreme loading, Journal of Ship Production & Design 31(3) (2018), 137–169.

21.

Dutta

A.K.

, Elhoseny

, Dahiya

and Shankar

, An efficient hierarchical clustering protocol for multihop Internet of vehicles communication, Transactions on Emerging Telecommunications Technologies (2019). In Press. DOI: 10.1002/ett.3690

22.

Elhoseny

, Shankar

and Uthayakumar

, Intelligent diagnostic prediction and classification system for chronic kidney disease, Nature Scientific Reports (2019). In Press. DOI: 10.1038/s41598-019-46074-2.

23.

Elhoseny

and Shankar

, Reliable data transmission model for mobile ad hoc network using signcryption technique, IEEE Transactions on Reliability (2019), 1–10. In Press. DOI: 10.1109/TR.2019.2915800

24.

Krishnaraj

, Elhoseny

, Thenmozhi

, Selim

M.M.

and Shankar

, Deep learning model for real-time image compression in Internet of Underwater Things (IoUT), Journal of Real-Time Image Processing (2019). In Press. DOI: 10.1007/s11554-019-00879-6.

25.

Elhoseny

and Shankar

, Optimal bilateral filter and convolutional neural network based denoising method of medical image measurements, Measurement 143 (2019), 125–135.

26.

Senthil Ragavan

V.K.

, Elhoseny

and Shankar

, An enhanced whale optimization algorithm for vehicular communication networks, International Journal of Communication Systems (2019). DOI: 10.1002/dac.3953

27.

Murugan

B.S.

, Elhoseny

, Shankar

and Uthayakumar

, Region-based scalable smart system for anomaly detection in pedestrian walkways, Computers and Electrical Engineering 75 (2019), 146–160.

28.

Shankar

, Elhoseny

, Eswaran Perumal

, Ilayaraja

and Kumar

, An efficient image encryption scheme based on signcryption technique with adaptive elephant herding optimization, cybersecurity and secure information systems: Challenges and solutions in smart environments, Advanced Sciences and Technologies for Security Applications, ISBN: 978-3-030-16837-7 (2019), 31–42.