KR101299391B1 - Scheduling using geneticalgorithm for maintenance in wind turbine equipment - Google Patents

Abstract

PURPOSE: A wind turbine maintenance scheduling method establishes an optimal scheduling for securing sufficient reserve power, reduces human errors, and decreases maintenance time and cost. CONSTITUTION: In terms of a gene's maintenance time, the genetic shape of initial parent groups is created within a normal distribution not to concentrate on a certain period of time. A share of the demand for reserve power for a wind turbine is calculated. A parent gene is selected according to the demand share. A breeding process or a mutation process is implemented regarding the parent gene. A maintenance schedule is developed as a result of the breeding process or the mutation process. [Reference numerals] (AA) Start; (BB) Define a period that requires maintenance for a whole wind turbine using a gene; (CC) Calculate the share (Y) of the annual maintenance period in a certain period of time (k); (DD) Is k/Y within a normal distribution ?; (EE) Add the result to a value group (initial parent group N); (FF) Define a reserve power value and a reserve power rate; (GG) Select a pair of chromosomes for the reserve power rate; (HH) Generate a first children chromosome by mating the parent chromosomes; (II) Is the reserve power of the first children larger than that of the parent ?; (JJ) Generate a second children chromosome by mutating the parent chromosomes; (KK) Is the first children or the second children fit for the maintenance period requirement ?; (LL) Add the first children or the second children to a children group; (MM) Is the number of the children group identical to the number (N) of the value group ?; (NN) Change the generation; (OO) Is a repetition number ending condition (M) satisfied ?; (PP) Evaluate an optimization rate; (QQ) End

Description

Scheduling using Genetic Algorithm for maintenance in Wind Turbine Equipment}

Wind Turbine Maintenance Scheduling

The theory is based on the genetic algorithm whose purpose is to select better chromosomes using gene chromosomes. Although it is widely used in modern science and industry, no maintenance algorithms are considered considering the unique characteristics of wind turbines.

 Power generation facilities, known as wind turbines, are a renewable energy industry that generates revenue by always producing electricity. After the facility is introduced, maintenance should be carried out to resolve the failures caused by the passage of time and the reduction in efficiency due to aging.

The problem is always a matter of order in which the maintenance should be performed according to the environment such as the condition of the equipment, the climate, and the required reserve power in the scheduling of the equipment for maintenance. One or two installations can be solved by the human brain, but at least the facility complexes above the heat can cause problems in human capacity. Even if scheduling is possible without human ability, it can bring a lot of trial and error and waste of time.

Therefore, this algorithm is an algorithm that can reduce the trial and error of human and waste of time and optimize the scheduling work of facility complex.

Chromosomes indicate when quarterly or monthly maintenance is required for a single turbine. This chromosome is extended by the total number of facilities in the power plant and is represented by one chromosome, S-expression.

The initial parent group constituting this chromosome should start with an even distribution of chromosomes for each period.

Securing reserve power is an important factor in maintenance, and as a factor for selecting each chromosome, the specific gravity according to the reserve power reserve of one chromosome should be used as a condition of fitness.

The chromosome selection is based on a high proportion of reserve power, which is called the parent chromosome. After mating and mutation, the chromosome is applied to the child chromosome group according to the degree of reserve power reserve and maintenance requirements.

The child's chromosome group repeats generational replacement with the parent's chromosome group, resulting in a better chromosome group.

Reserved power reserve rate is the most optimized scheduling task.

Reduce human trial and error and waste of time.

Reduce costs with optimized maintenance plans.

As a flowchart of a scheduling algorithm for maintenance of a wind turbine facility using a genetic algorithm,
The maintenance period required for wind turbine installations is defined as chromosomes and chromosome groups are created by specifying the standard deviation range so that the initial parent group can have various chromosomal groups.
In order to satisfy the goodness-of-fit conditions, the reserve reserve ratio is defined, and then the parent chromosome is selected according to the weight ratio of reserve reserve ratio.
The selected chromosome is a process of finding a good chromosome by performing crossbreeding and mutation to generate a group of children and repeating generations.

It is assumed that there are 10 facilities in the wind farm and the production power, load power and reserve power of each facility are as follows. Here, the load power may apply the same value to all 10 units.

34, 29, 5

2. 32, 28, 4

3. 29, 27, 2

4. 35, 28, 7

5. 33, 30, 3

6. 31, 28, 3

7. 29, 25, 4

8. 31, 26, 5

9. 32, 29, 3

10. 33, 31, 2

The period for maintenance in one phase of the installation is expressed in chromosomes. Suppose you need 2 or 3 maintenance during 12 months of the year

010001000000 or 000010100000. If you express it as 10 chromosomes

010001000000/000010100000/000010100100/000010001000/010000100000/100010100000/000010100001/000000010100/010100000000/000011100000. As many as N solution sets that require one such chromosome, the distribution of maintenance periods for each period should be within the standard deviation range so that they do not gather in a specific period.

The total power output of the 10 units is 319KW, the reserve power is 38KW.

1.5, 13%

2. 4, 11%

3. 2, 5%

4. 7, 18%

5. 3, 8%

6. 3, 8%

7. 4, 11%

8. 5, 13%

9. 3, 8%

10. 2, 5%

When selecting a chromosome from the initial parent population, it should be selected according to the proportion of reserve power reserves.

The child 1 chromosome generated by crossing over the selected parent chromosome is compared with the parent chromosome to secure reserve power. If the child 1 chromosome is high, the maintenance requirements are compared and placed in the child group. If the reserve power of the parent chromosome is high, the parent chromosome is mutated to make the child 2 chromosome, and then the maintenance conditions are compared and placed in the child group. If maintenance conditions are not satisfied or a child's chromosome enters a child group, the parent group will repeat the chromosome selection until the child set number N is reached.

When the number of child groups reaches N sets, the household is replaced, and the number of household replacements is repeated M times, and then ends with the optimization efficiency evaluation as the termination condition.

N: number of solution sets
M: Number of household changes
K: quarterly or monthly period
Y: Annual maintenance fraction ratio

Claims (6)

In the maintenance period selection method using genetic algorithm (GA) in the year-round maintenance schedule of the wind turbine installed in dozens,
In a period requiring maintenance, the first step of generating chromosomes within a range of a standard distribution so that the shape of the chromosomes of the first parent group is not biased at a specific time;
A second step of calculating a required reserve power ratio for one wind turbine facility;
Selecting a parent chromosome according to the required reserve power ratio;
A fourth step of performing crossing or mutation on the selected parent chromosome;
And a fifth step of determining a scheduling as a result of chromosomes for the performance.
The method of claim 1, wherein the first step,
Defining a time period in which the maintenance of each facility of the turbine is required as a chromosome; And
And determining the initial parent chromosome by determining whether the chromosome is included in the standard horseshoe range of each period.
The method of claim 1, wherein the second step,
A method of selecting a maintenance time using a genetic algorithm (GA) in a yearly maintenance schedule of a wind turbine, characterized in that the required reserve power of the wind turbine facility is calculated as the ratio of the value obtained by subtracting the power load from the power production capacity.
The method of claim 1, wherein the third step,
A method of selecting a maintenance time using a genetic algorithm (GA) in a yearly maintenance schedule of a wind turbine, characterized in that the parent chromosome is selected according to the proportion of reserve power required by each chromosome.
The method of claim 1, wherein the fourth step,
Crossing the parent chromosomes to generate a child 1 chromosome; And
Comparing the preliminary amount of power between the parent chromosome and the child 1 chromosome to determine the selection chromosome if the child 1 chromosome is large; And
If the parent chromosome is greater than the child chromosome in the selection, performing mutation to generate the child 2 chromosome and determining the selection chromosome; And
Determining whether the child chromosome and the child chromosome selected chromosome meet the maintenance requirement period; And
Determining the child group if the chromosome meets the maintenance requirement period; Method of selecting a maintenance time using a genetic algorithm (GA) in the yearly maintenance schedule of the wind turbine, characterized in that it comprises a.
The method of claim 1, wherein the fifth step,
Selecting the maintenance time using the genetic algorithm (GA) in the annual maintenance schedule of the wind turbine, characterized in that the step of applying to the schedule as a result of the chromosome when the determination of the number of sets of children and generation replacement is completed. Way.

Images