Maintenance and Operation Optimization Algorithm of PV Plants under Multiconstraint Conditions

Hua, Chi; Kuang, Liang; Pi, Dechang

doi:https://doi.org/10.1155/2020/7975952

Complexity

On this page

Abstract Introduction Data Availability Conflicts of Interest Acknowledgments References Copyright Related Articles

Special Issue

Deep Structure Representation and Learning for Complex Information Networks

View this Special Issue

Research Article | Open Access

Volume 2020 | Article ID 7975952 | https://doi.org/10.1155/2020/7975952

Maintenance and Operation Optimization Algorithm of PV Plants under Multiconstraint Conditions

Chi Hua,^1,2Liang Kuang,¹and Dechang Pi²

Academic Editor: Chuan Zhou

Received26 Jun 2020

Revised09 Aug 2020

Accepted05 Oct 2020

Published28 Oct 2020

Abstract

With the rapid increase in the photovoltaic (PV) plants, the real-time operation and maintenance of photovoltaic power generation equipment is very important. The maintenance and dispatching of decentralized power stations is still one of the key issues affecting the operation safety of photovoltaic power stations. However, most of the photovoltaic power stations in China fail to rationally optimize the utilization of resources and time. The current study puts forward effort implementation via genetic algorithm-based multiconstrained optimization methodology. The proposed study optimally overrides the traditional PV plant operation and maintenance dispatching operations with automation and reliability. The proposed study is also applicable to multiple PV plants, multiple maintainers, multipoint departure, different dispatching conditions, and cost considerations. We propose an MOOA algorithm to solve this issue, and we strongly believe that, by defining a suitable fitness function, the convergence speed and optimization ability can be greatly improved, and this study puts a forward step.

1. Introduction

Photovoltaic (PV) power plants are the core component of the smart grid [1, 2]. Because of its dispersed geographical location, complex structure, large number of equipment, and t complicated daily operation and maintenance procedures [3, 4], all these routine operations greatly affect the smooth operation of the smart grid. At present, the maintenance of PV plants [5–7] is mainly carried out in accordance with the regular maintenance stipulated by the state. There exist a number of problems which lead to excessive cost maintenance, such as insufficient maintenance, poor reliability, wasting of resources, and increasing maintenance costs. Therefore, operation and maintenance dispatch has become the core problem of an intelligent operation system of PV plants [8, 9]. Operation and maintenance assignment is essentially an extension of the assignment problem. The assignment of tasks to persons belongs to balanced assignment; otherwise, it belongs to an unbalanced assignment [10]. In the process of operation and maintenance of PV plants, the number of maintainers is less than that of power plants, which leads to the unbalanced assignment problem. Balanced assignment problem can be well solved by the Hungarian algorithm, eliminating heights and shrinking matrix analysis [11], while unbalanced assignment problem is an objective function optimization problem with constraints, and constraints and objective functions may not be linearized, so it is difficult to solve the problem. A couple of research studies have put forward efforts to solve an unbalanced assignment problem [12, 13], such as dual transportation method [14–17] and fuzzy Hungarian algorithm [18]. However, the solving process is complex and does not have robustness and adaptability. The machine learning especially deep learning also provides a solution to the unbalanced assignment problem [19]. In this paper, a genetic algorithm is presented to solve the unbalanced assignment problem of operation and maintenance dispatch. This method can optimize the operation and maintenance dispatch of multiple PV plants, multiple maintainers, and multipoint departure. It can set different assignment constraints or cost considerations and accelerate the convergence speed and improve the optimization ability of the algorithm.

2. Establishment of Operational and Maintenance Model of PV Plants

The assignment problem takes the matching of tasks and maintainers as the research object to optimize the utilization of resources [20, 21]. The mathematical model of operation and maintenance dispatch for PV plants is described as follows.

Assuming that the number of maintainers is and the number of tasks is at a certain time, it obtains by considering the maintenance cost, and one plant is only maintained by one maintainer; the maintainers will perform the next task according to the algorithm hints after the accomplishment of last PV plant’s maintenance. In this paper, the optimal path of operation and maintenance dispatch is given to minimize the time spent by maintainers on the road. represents the time consuming on the road between two maintained PV plants, i.e.,

Then, the mathematical model of multiobjective unbalanced assignment for maintainers to PV plants is as follows: where means that the maintainer completes the task and means that no task is assigned.

In order to solve the unbalanced assignment in (2), it makes an equivalent exchange for (2), that is, maintainers are added to make up maintainers. In essence, the added maintainers are only considered in the calculation and will not be shown in dispatch route planning; moreover, the added maintainers will not affect the objective function and constraints, so the essence of the unbalanced assignment problem has not changed. The original unbalanced assignment problem is equivalent to the balanced assignment problem, that is, m maintainers are responsible for the operation and maintenance of n PV plants. The mathematical model is as follows:where attributing 0-1 variables correspond to the task assignment of adding maintainers. Obviously, the transformed model is equivalent to the original model. Equation (4) only corresponds to one assignment, which satisfies both constraints and objective optimization. Traditional methods are difficult to solve it. Essentially, the assignment problem of the operation and maintenance of PV plants is an NP-complete problem, which is solved by searching one by one, and time consumed will be . In this paper, we have employed a genetic algorithm to solve the operation and maintenance dispatch of PV plants.

3. Optimization Algorithm of Maintainers Dispatching

The bases of the genetic algorithm for solving the optimal maintenance strategy of PV plants are as follows: First, the genetic algorithm solves various TSP problems with many results [22–24] and is mature and reliable. Second, the genetic algorithm satisfies the solution of different problems; for instances, the problem of multipoint departure and multiperson operation and maintenance can be assigned by adjusting the chromosome coding method, and the operation and maintenance strategy with uncertain time constraints and cost calculation methods can be made by adjusting the fitness function. Third, the genetic algorithm is easy to iterate, which meets the new demands emerging from the operation and maintenance environment of actual PV plants [25, 26].

As shown in Figure 1, the fault diagnosis process of PV array is described. This paper designs the optimal distribution of voltage sensors according to the distribution of PV plants by equivalent analysis of the weight matrix. Through regression analysis of state data [27, 28], the location of the fault point is obtained by analyzing the voltage signal detected via the voltage sensor. The fault diagnosis model of PV arrays is designed and carried out. This model classifies the types of fault points [29–31]. Based on the concept of equipment lifecycle management [32, 33], this paper establishes the equipment status data model, obtains the information of equipment management, establishes the health history of equipment status, refers to the provisions of equipment manufacturers on their design life, effective operation life, maximum allowable life, and the sample statistical data of their actual service life, and achieves the assistant decision making of equipment condition monitoring in photovoltaic power plants.

The following steps are used to optimize the operation and maintenance dispatch of PV plants: Step 1. Input the fault information to be solved, including the number of faults , the location of the faulty power station , the installed capacity of the faulty power station , the fault grade of the faulty power station , the remaining maintenance waiting time of the faulty power station , and the fault flow number . Step 2. The path optimization algorithm is obtained by the genetic algorithm.①Input the traffic cost matrix, that is, the traffic cost between the fault location and the departure location.②Random sequencing of fault pipeline numbers constitutes the initial gene as follows:③Repeat ② for times to form initial genes: ; initial genes constitute the initial population .④Construct fitness function : where represents the gene of fitness function to evaluate fitness, represents the fault flow number in the gene, represents the traffic cost between the two fault locations determined by the fault flow number and the traffic cost matrix C, represents the total traffic cost of maintenance in the order of the intermediate flow number, represents the order of in ; , if it ranks first in . denotes the installed capacity of the faulty power station corresponding to the fault pipeline number , denotes the fault grade of the faulty plant corresponding to the fault pipeline number , and denotes the remaining maintenance waiting time of the faulty power station corresponding to the fault pipeline number . represents the penalty coefficient of the traffic cost; the default is 1000; the larger m is, the lower the route of the traffic cost will be preferred, and represents the penalty coefficient of the installed capacity; the default is 100; the larger is, the preferred choice will be made. For a power station with large installed capacity, means the penalty coefficient of the fault grade. The bigger is, the higher the penalty coefficient of maintenance time will be. means the penalty coefficient of maintenance time. The bigger is, the lower the waiting time of maintenance will be.⑤ individuals were randomly selected from the population , and the probability of individual being selected was as follows: that is, the greater the fitness, the higher the probability of being selected, and the new individuals selected are inserted into the new population .⑥The probability is used to randomly select individuals from the population , which is even and convenient for crossover operation. The crossover operation is carried out by randomly extracting a number of individual gene fragments and pairing individuals to form a new individual to join the population . In the optimization process of operation and maintenance of PV plants, is set to 0.75. In the new population , the probability is used to randomly select individuals, which is for mutation. That is to say, two gene fragments are randomly selected, and their positions are exchanged to form new individuals to join the population . In the optimization process of operation and maintenance of PV plants, is set to 0.01. The crossover method and the mutation method are shown in Figure 2. Two individuals were selected randomly with probability to perform crossover operation. The offspring inherited part of the genes from their parents and kept the same sequence as their parents are as shown in Figure 2. Some genes of Parent 1, such as 678, crossed with Parent 2 were selected randomly, and the crossover results are shown in Figure 2. The probability of mutation occurs. Because each city only passes through once in TSP problem, the mutation should not change the value of one position in the gene sequence (which will cause a city to pass through twice) but should exchange the values of two positions randomly, as shown in Figure 2; the positions of 3 and 8 should be exchanged.⑦Repeat ⑤ and ⑥ until the maximum number of iterations is reached, or the fitness function of an individual reaches a given threshold.⑧For the individual with the largest output fitness, the sequence of flow numbers is the maintenance sequence given by the algorithm. Step 3. The optimal operation and maintenance scheme is output by the path optimization algorithm. Step 4. If when the maintainers implement the optimal path, the operation and maintenance time consuming does not match the expected time consuming or new faults occur during the operation and maintenance period; we need to update the information to solve the fault and repeat Steps 2–4 again, until the maintainers successfully implement the optimal path.

4. Experiments and Implementation Verification

The experimental environment uses MATLAB 6.0 to construct 36 random points as the distribution of PV plants in the region and triangulate the points, and the time spent on the road is the sum of the edge lengths of the triangle experienced between the two points; construct the weight matrix of the operation and maintenance process, randomly select a number of points from the points as maintenance objects, and set them up again every 1 minute from the points. Randomly select 0 to 1 point as a new maintenance object to join the maintenance queue. Suppose a maintainer is at a fixed speed between the set of points in the operation and maintenance, and the speed per minute does not exceed the length of the maximum edge of the triangle. As shown in Figure 3, randomly generate points, triangulate points, and construct maintenance site.

In the experiment, the maintenance of PV plants can be well carried out by implementing one maintainer and three maintainers, respectively. By implementing a genetic algorithm, the optimal dispatching path of maintainers can be saved, especially when facing 38 photovoltaic power stations, no more than three maintainers can be maintained.

In order to better prove the rationality of the optimization method of operation and maintenance dispatch of PV plants based on genetic algorithm, random method, greedy algorithm, and genetic algorithm were used to carry out experiments. The time consuming on the road between two places of a company’s PV plants in Zhejiang province is shown in Table 1, including Hangzhou, Deqing, Haining, Huzhou, Jiande, Kaihua, Kecheng, Linan, Linhai, Longquan, Pan’an, Quzhou, Shangyu, Shaoxing, Shengzhou, Taizhou, Tiantai, Tonglu, Tongxiang, Yuyao, and Zhuji. The PV plants in all areas are repaired and maintained once to find an appropriate operation and maintenance path. The average travel time for each region is 159 minutes, of which(1)The average total time of 21 horizons patrol by the random route method is 3340 minutes.(2)Using a greedy algorithm (the next destination is the nearest location from the current location), the route is selected as follows: Hangzhou > Tongxiang > Haining > Shangyu > Shaoxing > Yuyao > Shengzhou > Tiantai > Linhai > Taizhou > Pan’an > Zhuji > Linan > Deqing > Huzhou > Tonglu > Jiande > Kecheng > Quzhou > Kaihua > Longquan > Hangzhou. The time series is {61, 35, 65, 45, 40, 86, 69, 54, 68, 144, 99, 98, 73, 58, 140, 66, 71, 13, 61, 217, 313}, totaling 1876 minutes.(3)The genetic algorithm is used to optimize the maintenance path. The time cost matrix of operation and maintenance location is shown in Table 1. The selected route is as follows: Hangzhou > Linan > Deqing > Huzhou > Tongxiang > Haining > Zhuji > Shaoxing > Yuyao > Shangyu > Shengzhou > Tiantai > Taizhou > Linhai > Pan’an > Longquan > Kaihua > Kecheng > Quzhou > Jiande > Tonglu > Hangzhou.

The time series is {62, 73, 58, 87, 35, 103, 77, 40, 55, 67, 69, 90, 68, 97, 187, 217, 56, 13, 76, 66, 78}, totaling 1278 minutes.

Compared with the optimal path obtained by the exhaustive wet method, it is confirmed that the improved genetic algorithm can obtain the optimal path for 21 operation and maintenance routes.

The coding method of the genetic algorithm and the definition of fitness function in [34, 35] were used to solve the operation and maintenance route planning of PV plants. Comparing with this algorithm, as shown in Figure 4, [35] needs 81 iterations, [34] needs 44 iterations, while this algorithm only needs 34 iterations, and the final operation and maintenance path takes 1278 minutes. By defining an appropriate fitness function, the strategy of a genetic algorithm to optimize the operation and maintenance path presented in this paper can accelerate the convergence speed and searchability of the genetic algorithm.

5. Summary

This paper presents an optimization method of operation and maintenance dispatch of PV plants based on the genetic algorithm. This method aims at the operation and maintenance tasks of large-scale distributed PV plants. GA for M-TSP is applied to provide an optimal path for maintainers dispatched for PV plants. It supports the optimization of operation and maintenance dispatch of multiple PV plants, multimaintainer, multiobjective, and multipoint departure. By multitype cost function, different assignment constraints or cost considerations can be set, such as arrival within a specified time or minimum time cost of transportation costs. We ascertain that the proposed model will lead the way.

Data Availability

The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest

The authors declare that they have no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Acknowledgments

Some of the authors of this publication are also working on these related projects: (1) Higher Vocational Education Teaching Fusion Production Integration Platform Construction Projects of Jiangsu Province under grant no. 2019(26), (2) Natural Science Fund of Jiangsu Province under grant no. BK20131097, (3) “Qing Lan Project” Teaching Team in Colleges and Universities of Jiangsu Province under grant no. 2017(15), (4) High Level of Jiangsu Province Key Construction Project funding under grant no. 2017(17), and (5) Natural Science Research in Jiangsu Universities in 2020.

References

S. Prakesh, S. Sherine, and B. Bist, “Forecasting methodologies of solar resource and PV power for smart grid energy management,” International Journal of Pure and Applied Mathematics, vol. 116, no. 18, pp. 313–318, 2017.
View at: Google Scholar
M. S. Hossain, N. A. Madlool, N. A. Rahim, J. Selvaraj, A. K. Pandey, and A. F. Khan, “Role of smart grid in renewable energy: an overview,” Renewable and Sustainable Energy Reviews, vol. 60, pp. 1168–1184, 2016.
View at: Publisher Site | Google Scholar
S. S. S. R. Selvaraj, L. Wang, V. Devabhaktuni, and N. Gudi, “Smart meters for power grid—challenges, issues, advantages and status,” in Proceedings of the IEEE/PES Power Systems Conference and Exposition, pp. 1–7, Phoenix, AZ, USA, March 2011.
View at: Publisher Site | Google Scholar
Y. Kabalci, “A survey on smart metering and smart grid communication,” Renewable and Sustainable Energy Reviews, vol. 57, pp. 302–318, 2016.
View at: Publisher Site | Google Scholar
M. De Benedetti, F. Leonardi, F. Messina, C. Santoro, and A. Vasilakos, “Anomaly detection and predictive maintenance for photovoltaic systems,” Neurocomputing, vol. 310, no. 1, pp. 59–68, 2018.
View at: Publisher Site | Google Scholar
V. Mauro, C. Vittorio, L. Alfonsi, and V. Introna, “Optimization of photovoltaic maintenance plan by means of a FMEA approach based on real data,” Energy Conversion and Management, vol. 152, no. 1, pp. 1–12, 2017.
View at: Publisher Site | Google Scholar
M. Di Somma, B. Yan, N. Bianco et al., “Multi-objective design optimization of distributed-energy systems through cost and exergy assessments,” Applied Energy, vol. 204, no. 1, pp. 1299–1316, 2017.
View at: Publisher Site | Google Scholar
S. Affijulla and S. Chauhan, “A new intelligence solution for power system economic load dispatch,” in Proceedings of the 10th International Conference on Environment and Electrical Engineering, pp. 1–5, Rome, Italy, May 2011.
View at: Publisher Site | Google Scholar
J. Gao, Y. Xiao, J. Liu, W. Liang, and C. L. P. Chen, “A survey of communication/networking in Smart Grids,” Future Generation Computer Systems, vol. 28, no. 2, pp. 391–404, 2012.
View at: Publisher Site | Google Scholar
W. Liang, D. Pi, and Z. Shao, “A pareto-based estimation of distribution algorithm for solving multiobjective distributed No-wait flow-shop scheduling problem with sequence-dependent setup time,” IEEE Transactions on Automation Science and Engineering, vol. 16, no. 3, pp. 1344–1360, 2019.
View at: Publisher Site | Google Scholar
S. H. I. Zhong-min, “General assignment problem,” Operations Research and Management Science, vol. 8, no. 1, pp. 21–26, 1999.
View at: Google Scholar
Y. Song, M.-Y. Chen, and Z. Shu-hong, “Multi-objective decision-making method for site selecting of projects based on fuzzy assignment,” System Engineering—Theory and Practice, vol. 22, no. 1, pp. 108–111, 2002.
View at: Google Scholar
H. De-cai, “An efficiency algorithm for solving the optimal solution of a generalized assignment problem,” Control and Decision, vol. 14, no. 3, pp. 272–275, 1999.
View at: Google Scholar
F. Glover, D. Karney, and D. Klingman, “A computation study on start procedures, basis change criteria, and solution algorithms for transportation problems,” Management Science, vol. 20, no. 5, pp. 793–813, 1974.
View at: Publisher Site | Google Scholar
S. Napier, “General assignment problem,” Operations Research and Management Science, vol. 1, 1999.
View at: Google Scholar
A. Kumar, S. S. Appadoo, and P. Kaur, “Mehar approach for solving dual-hesitant fuzzy transportation problem with restrictions,” Sādhanā, vol. 45, no. 1, pp. 1–9, 2020.
View at: Publisher Site | Google Scholar
L. Cui, J. Wu, D. Pi, P. Zhang, and P. Kennedy, “Dual implicit mining-based latent friend recommendation,” IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 50, no. 5, pp. 1663–1678, 2018.
View at: Publisher Site | Google Scholar
X.-Z. QinW. Xue-hua, ““Dynamic programming model of a sort of optimal assignment problem,” Mathematics in Practice and Theory, vol. 26, no. 3, pp. 212–216, 1996.
View at: Google Scholar
F. Liu, S. Xue, J. Wu et al., “Deep Learning for Community Detection: Progress, Challenges and Opportunities,” in Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence, Yokoham, Japan, July 2020.
View at: Publisher Site | Google Scholar
B. Ming, P. Liu, S. Guo, X. Zhang, M. Feng, and X. Wang, “Optimizing utility-scale photovoltaic power generation for integration into a hydropower reservoir by incorporating long- and short-term operational decisions,” Applied Energy, vol. 204, pp. 432–445, 2017.
View at: Publisher Site | Google Scholar
F.-F. Zhang and J. Qiu, “Multi-objective optimization for integrated hydro-photovoltaic power system,” Applied Energy, vol. 167, pp. 377–384, 2016.
View at: Publisher Site | Google Scholar
J. Yan-Qun, Z. Shu-Guang, and C. Bokui, “Fixed-point fast sweeping weighted essentially non-oscillatory method for multi-commodity continuum traffic equilibrium assignment problem,” Applied Mathematical Modelling, vol. 62, pp. 404–414, 2018.
View at: Publisher Site | Google Scholar
Z. Wang, P. Jun, C. Liling et al., “A parallel biological optimization algorithm to solve the unbalanced assignment problem based on DNA molecular computing,” International Journal of Molecular Sciences, vol. 16, no. 10, pp. 25338–25352, 2015.
View at: Google Scholar
S.-U. Lee, “Assignment problem algorithm using the dual-path competition method,” Journal of Korean Institute of Information Technology, vol. 13, no. 12, pp. 75–83, 2015.
View at: Publisher Site | Google Scholar
P. A. Kumari and P. Geethanjali, “Parameter estimation for photovoltaic system under normal and partial shading conditions: a survey,” Renewable and Sustainable Energy Reviews, vol. 84, pp. 1–11, 2018.
View at: Publisher Site | Google Scholar
T. S. Babu, J. P. Ram, T. Dragičević et al., “Particle swarm optimization based solar PV array reconfiguration of the maximum power extraction under partial shading conditions,” IEEE Transactions on Sustainable Energy, vol. 9, no. 1, pp. 74–85, 2018.
View at: Publisher Site | Google Scholar
C. A. MacKenzie and K. Barker, “Empirical data and regression analysis for estimation of infrastructure resilience with application to electric power outages,” Journal of Infrastructure Systems, vol. 19, no. 1, pp. 25–35, 2013.
View at: Publisher Site | Google Scholar
M. Hassanzadeh and C. Y. Evrenosoglu, “A regression analysis based state transition model for power system dynamic state estimation,” in Proceedings of the 2011 North American Power Symposium, Boston, MA, USA, August 2011.
View at: Publisher Site | Google Scholar
J. Wu, S. Pan, X. Zhu et al., “Boosting for multi-graph classification,” IEEE Transactions on Cybernetics, vol. 45, no. 3, pp. 416–429, 2014.
View at: Publisher Site | Google Scholar
J. Wu, X. Zhu, C. Zhang, and P. S. Yu, “Bag constrained structure pattern mining for multi-graph classification,” IEEE Transactions on Knowledge and Data Engineering, vol. 26, no. 10, pp. 2382–2396, 2014.
View at: Publisher Site | Google Scholar
C. Yu, D. Pi, S. I. Becker, L. Cui, and B. Johnson, “CenEEGs,” ACM Transactions on Knowledge Discovery from Data, vol. 14, no. 2, pp. 1–25, 2020.
View at: Publisher Site | Google Scholar
D. Wu, A. Bufardi, and P. Xirouchakis, “Research issues on product lifecycle management and information tracking using smart embedded systems,” Advanced Engineering Informatics, vol. 17, no. 3-4, pp. 189–202, 2003.
View at: Publisher Site | Google Scholar
L. A. I. Yun and L. U. Chen, “Construction of entire lifecycle management model of laboratory equipment in universities,” Research and Exploration in Laboratory, vol. 2, 2012.
View at: Google Scholar
I. Younas, F. Kamrani, M. Bashir et al., “Efficient genetic algorithms for optimal assignment of tasks to teams of agents,” Neurocomputing, vol. 314, pp. 409–428, 2018.
View at: Publisher Site | Google Scholar
B. Marin, D. Guillerme, and G. Rodolphe, “Approximability and exact resolution of the multidimensional binary vector assignment problem,” Journal of Combinatorial Optimization, vol. 36, no. 3, pp. 1059–1073, 2018.
View at: Publisher Site | Google Scholar

Copyright

Copyright © 2020 Chi Hua et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

PDF Download Citation

Download other formats

Order printed copies

Views

355

Downloads

845

Citations