Data evaluation method for ceramic 3D printing samples based on fuzzy algorithm

Abstract

In order to improve the evaluation score and accuracy of ceramic 3D printing sample data, a fuzzy algorithm based evaluation method of ceramic 3D printing sample data is proposed. The data of ceramic 3D printing samples were collected and preprocessed by fuzzy algorithm. The specific content of the evaluation index of 3D printing sample data is determined, and the evaluation system of 3D printing sample data is established. Fuzzy algorithm is used to determine the weight of data quality evaluation index of ceramic 3D printing samples. According to the steps of setting up the weight matrix of the evaluation index of 3D printed ceramic sample data, the weight matrix of the evaluation index is established. Because there are many data sources of ceramic 3D printing samples, there may be data conflicting with the description of the same object. Under the background of fuzzy algorithm, the effective dimension of ceramic 3D printing sample data is established. According to the membership function of each evaluation index, the fuzzy set of 3D printing sample data of ceramics is established. Finally, the data evaluation process is designed to realize the data evaluation of ceramic 3D printing samples based on fuzzy algorithm. The experimental results show that the evaluation accuracy of the method is higher than 96%, and the scores of users and experts on the interior design scheme are above 80 points.

Keywords

Fuzzy algorithm ceramic material 3D printing sample data data evaluation

1 Introduction

China’s ceramics has a long history and has long been known in various countries around the world. The president of Germany’s Hendry Group once said that China is not only the place with the oldest ceramics, but also the place with the richest ceramic traditions [18]. China is known as a ceramic power, but it cannot be called a ceramic power. This situation can be attributed to a variety of reasons, one of the important reasons is from data evaluation for ceramic 3D printing samples.

Ceramic products have the advantages of high temperature resistance, corrosion resistance, high hardness and high strength, which are widely used in aerospace, machinery, electronics and other fields. The research shows that 3D printing sample data evaluation is the key factor to reduce the production cost of ceramic products and improve the molding performance of ceramic parts. The production process of traditional ceramic products includes three steps of raw material process, forming process and firing process [19], the most complicated process is forming process, and the data evaluation of this process is also the most difficult. As the ceramic products are produced by molds, the surface quality of the products will be reduced, and subsequent operations such as 3D sample data evaluation are required. From this, we can see that the traditional evaluation has the following shortcomings: low efficiency, easy to cause data loss, high cost, long evaluation time, etc., which cannot meet the needs of the market and the development of science and technology.

Evaluation method is also called “additive manufacturing data evaluation method”. Because of its influence on ceramic industry, it is also called “the third industrial revolution”. It is totally different from the traditional evaluation method. This method uses computer 3D modeling and fuzzy algorithm to evaluate 3D sample data, so it can speed up the efficiency of data evaluation. Fuzzy algorithm is a kind of intelligent algorithm which was proposed by American scientists in 1960 s. Classification evaluation is an important part of fuzzy algorithm research. Its corresponding classification system includes parallel generation subsystem based on string rules, fuzzy algorithm subsystem and data evaluation subsystem. It has been widely used in economy, science and engineering. Common fuzzy algorithms, such as mean blur, Gaussian blur and so on, the basic process is to calculate the sum of some eigenvalues and their corresponding weights in a certain field around a pixel, and then get the result value. From the operation process of the fuzzy algorithm, this method belongs to the intelligent algorithm, but strictly speaking, it is not a simple intelligent fuzzy. The algorithm can optimize the algorithm results by effectively using the existing relevant information to improve the adaptability of the data. Similar to the principle of survival of the fittest in nature, fuzzy algorithm cannot know the problem solving process in advance. The information that the algorithm relies on mainly comes from the evaluation results of each 3D ceramic sample data. The weight of 3D ceramic sample data is determined by the corresponding fitness value, so that the information of 3D ceramic sample data with good fitness value can be better saved [21]. Through ingenious use of coding technology and fuzzy algorithm theory, we can effectively solve some complex phenomena and difficult problems in the evaluation of 3D ceramic sample data.

With the development of science and technology, many domestic and foreign enterprises are actively exploring how to make good use of 3D sample data to keep the ceramic industry in an invincible position in the fierce social competition, and even to find and open up new high added value fields. However, due to the unbalanced development of economic system or society, some domestic ceramic enterprises are aware of the importance of 3D sample data to the production and operation management of enterprises, but they do not know how to use 3D sample data or cannot effectively use 3D sample data. The definition of 3D sample data usually exists in a narrow sense and a broad sense. The 3D sample data in this paper refers to the sample data generated in the production process of ceramic enterprises. These sample data include not only the basic information of enterprises, but also the operation performance of enterprises. Various problems found in the investigation of some enterprises. Considering the problems of too high dimension and too large data of 3D sample data, this paper first eliminates the incompleteness of 3D sample data and the correlation between dimensions, so that the data can meet the use requirements for the subsequent data processing and data application. Secondly, in view of the specific problems of ceramic enterprises, this paper studies the corresponding data evaluation system, and compares the advantages and disadvantages of the evaluation indexes. From the full text, this paper focuses on how to deal with 3D sample data and how to apply 3D sample data. Through the processing and application of 3D sample data, it provides corresponding support for the operation management and decision-making of ceramic enterprises [1].

3D sample data is the basis of analysis and decision-making of ceramic enterprises, and data quality is related to the use of data, so data quality plays an irreplaceable key role in the effectiveness and scientificity of management. Evaluation is one of the most important methods to solve data quality problems. Although the concept of data quality is different, it is generally believed that data quality can be divided into different categories and different evaluation dimensions according to different applications [17]. The core of data quality evaluation is how to evaluate each dimension concretely. For a long time, many scholars dimensions to the evaluation of data, carried on the thorough research, but most of the effect is not ideal, therefore low evaluation score and evaluation the problem of poor accuracy, in order to solve the traditional methods evaluation score is low, and evaluate the problem of low accuracy, so this article will solve the problem, as the key to 3D printing ceramic sample data to evaluate the goal of high grade and high accuracy, is proposed based on fuzzy algorithm of ceramic samples of 3D printing data evaluation method. This article study, usually adopt the data accuracy, timeliness, integrity, objectivity, availability, dimensions such as useful to measure the quality of the data, can effectively promote the ceramic and accuracy evaluation score, 3D printing sample data for the sample data evaluation method lays the foundation for the further development, and promote the further development in the field of ceramic 3D printing.

2 Materials and methods

This paper takes data for ceramic 3D printing samples as the research object, introduces the commonly used ceramic 3D printing technology, and analyzes the temporal and spatial characteristics and internal correlation of data for ceramic 3D printing samples. According to the theory and characteristics of 3D printing, the quality of data for ceramic 3D printing samples was analyzed. According to the logical order of normative detection, missing data detection, wrong data and inaccurate data detection [9], this paper puts forward the evaluation method of data for ceramic 3D printing samples. Based on the quality analysis of data for ceramic 3D printing samples, the quality evaluation index system of data for ceramic 3D printing samples is proposed from micro, meso and macro levels. Fuzzy algorithm is used to evaluate the quality of data for ceramic 3D printing samples. Finally, on the basis of quality problem identification and evaluation of data for ceramic 3D printing samples, the fuzzy set of data for ceramic 3D printing samples is built to improve the quality of data for ceramic 3D printing samples. The technical roadmap of this paper is shown in Fig. 1.

Fig. 1

Technology roadmap.

2.1 Sample data preprocessing based on fuzzy algorithm

Data for ceramic 3D printing samples mainly comes from some data in a ceramic enterprise database, involving 84 data tables, including customer information table, business code table, detailed business table, user status table, upload file record table, evaluation dimension table, evaluation data table, rule table, etc. The data contained in the corresponding data table is shown in Table 1.

Table 1
Corresponding table of ceramic 3D printing sample data

Code Data volume Code Data volume Code Data volume Code Data volume Code Data volume

Surface 1 1849488 Surface 18 956 Surface 35 734079 Surface 52 1474720 Surface 69 576321

Surface 2 2940942 Surface 19 2033 Surface 36 732032 Surface 53 5311209 Surface 70 552316

Surface 3 1032506 Surface 20 2973171 Surface 37 732226 Surface 54 5252359 Surface 71 523647

Surface 4 1979137 Surface 21 442 Surface 38 723509 Surface 55 5348132 Surface 72 510347

Surface 5 3879651 Surface 22 1562042 Surface 39 744274 Surface 56 5263237 Surface 72 562741

Surface 6 3590 Surface 23 24 Surface 40 742688 Surface 57 5108584 Surface 74 520146

Surface 7 13 Surface 24 14645148 Surface 41 743496 Surface 58 5195895 Surface 75 591463

Surface 8 46 Surface 25 776778 Surface 42 742628 Surface 59 5203739 Surface 76 5436751

Surface 9 258 Surface 26 731443 Surface 43 745069 Surface 60 5526673 Surface 77 5103246

Surface 10 693072 Surface 27 761023 Surface 44 742628 Surface 61 552578 Surface 78 632548

Surface 11 786278 Surface 28 781701 Surface 45 745069 Surface 62 536537 Surface 79 603542

Surface 12 1927577 Surface 29 787945 Surface 46 749056 Surface 63 533246 Surface 80 572384

Surface 13 33 Surface 30 773766 Surface 47 764186 Surface 64 532648 Surface 81 584167

Surface 14 46605 Surface 31 767055 Surface 48 181322 Surface 65 554982 Surface 82 501432

Surface 15 186025 Surface 32 724338 Surface 49 144957 Surface 66 523469 Surface 83 542168

Surface 16 23595 Surface 33 731399 Surface 50 151561 Surface 67 554195 Surface 84 585837

Surface 17 7663280 Surface 34 732679 Surface 51 146564 Surface 68 510940

Code	Data volume	Code	Data volume	Code	Data volume	Code	Data volume	Code	Data volume
Surface 1	1849488	Surface 18	956	Surface 35	734079	Surface 52	1474720	Surface 69	576321
Surface 2	2940942	Surface 19	2033	Surface 36	732032	Surface 53	5311209	Surface 70	552316
Surface 3	1032506	Surface 20	2973171	Surface 37	732226	Surface 54	5252359	Surface 71	523647
Surface 4	1979137	Surface 21	442	Surface 38	723509	Surface 55	5348132	Surface 72	510347
Surface 5	3879651	Surface 22	1562042	Surface 39	744274	Surface 56	5263237	Surface 72	562741
Surface 6	3590	Surface 23	24	Surface 40	742688	Surface 57	5108584	Surface 74	520146
Surface 7	13	Surface 24	14645148	Surface 41	743496	Surface 58	5195895	Surface 75	591463
Surface 8	46	Surface 25	776778	Surface 42	742628	Surface 59	5203739	Surface 76	5436751
Surface 9	258	Surface 26	731443	Surface 43	745069	Surface 60	5526673	Surface 77	5103246
Surface 10	693072	Surface 27	761023	Surface 44	742628	Surface 61	552578	Surface 78	632548
Surface 11	786278	Surface 28	781701	Surface 45	745069	Surface 62	536537	Surface 79	603542
Surface 12	1927577	Surface 29	787945	Surface 46	749056	Surface 63	533246	Surface 80	572384
Surface 13	33	Surface 30	773766	Surface 47	764186	Surface 64	532648	Surface 81	584167
Surface 14	46605	Surface 31	767055	Surface 48	181322	Surface 65	554982	Surface 82	501432
Surface 15	186025	Surface 32	724338	Surface 49	144957	Surface 66	523469	Surface 83	542168
Surface 16	23595	Surface 33	731399	Surface 50	151561	Surface 67	554195	Surface 84	585837
Surface 17	7663280	Surface 34	732679	Surface 51	146564	Surface 68	510940

These 3D printing sample data tables have a lot of noise information without preprocessing. In addition, different application targets need different 3D printing sample data to support [20]. Therefore, this section first deals with and filters the 3D printing sample data. First of all, in order to analyze the effectiveness, starting from the basic information of 3D printing sample data, 3D printing sample data with incomplete information and wrong information are screened, so that the 3D printing sample data used in this section has complete information. Then the basic information of 3D printing sample data is used to add the format information of 3D printing sample data. Thirdly, according to the network access time of 3D printing sample data, the corresponding network time of 3D printing sample data is calculated. Fourth, according to the status information of 3D printing sample data and the 3D printing sample data information used, the relevant data are screened. Fifth, according to the business code of 3D printing sample data, all kinds of business information of 3D printing sample data are separated and summarized, and the evaluation trend data of 3D printing sample data is calculated [7]. According to the above data preprocessing and other statistical work, a data set containing basic information and various evaluation information of 3D printing sample data is finally obtained. From this data set, 8000 pieces of data are randomly selected as the training set of 3D printing sample data evaluation, and 20000 pieces of data are selected as the evaluation test set.

The preprocessing process of ceramic 3D printing sample data is shown in Fig. 2.

Fig. 2

Preprocessing flowchart of ceramic 3D printing sample data.

Data preprocessing can exist in all aspects of data evaluation. By predicting the trend of the market, the needs of customers, and controlling the cost of enterprises, ceramic enterprises can optimize their business management activities [10]. The effect of preprocessing depends partly on data selection and corresponding data collection, while the greater part is affected by preprocessing methods. In this paper, through the analysis of different problems, the corresponding data set is selected and the corresponding evaluation system is constructed. Through the study of different problems, the common purpose of cost control and management optimization of ceramic enterprises is finally achieved.

2.2 Construction of 3D printing sample data evaluation system for ceramics

At present, there is no uniform specification or standard for the evaluation of data for ceramic 3D printing samples at home and abroad, and different types of data for ceramic 3D printing samples will have different data evaluation systems [24]. According to the preprocessing principle of data for ceramic 3D printing samples, the evaluation system of data for ceramic 3D printing samples is established as shown in Fig. 3.

Fig. 3

Evaluation systems of data for ceramic 3D printing samples.

where U₁, U₂, ⋯ U₉ respectively represent the evaluation index sets of different types of data for ceramic 3D printing samples. The specific meanings are as follows:

$\begin{matrix} Q_{1} = {U_{1}, U_{2}} = \\ {Normalization, Property integrity} \end{matrix}$ (1)

$\begin{matrix} Q_{2} = {U_{3}, U_{4}, U_{5}, U_{6}, U_{7}} = \\ {Accuracy, Effectiveness, timeliness, \\ Content integrity, Coverage} \end{matrix}$ (2)

$Q_{3} = {U_{8}, U_{9}} = {Creditability, Usability}$ (3)

The specific contents of the evaluation indexes of data for ceramic 3D printing samples are shown in Table 2.

Table 2

Specific contents of evaluation indexes of data for ceramic 3D printing samples

Target	First level index	Two level index	Concrete content
Evaluation of 3D printing sample data of ceramics	Quality of data presentation Q1	Normalization	Format normalization
		Property integrity	Whether attribute information is missing
	Data content quality Q2	Accuracy	Relative error with correct value
		Effectiveness	Is the data within a reasonable threshold
			Whether it conforms to the three parameter relationship
			Data to determine whether there is discrete rule test
			Judge continuous same record inspection
		Timeliness	Data update synchronization time
		Content integrity	Data integrity
		Coverage	For a single detector, coverage is the presence or absence of data
			For medium or macro data quality evaluation
	Data processing quality Q3	Creditability	Data credibility compared with previous experience
		Usability	Whether it is of great help to the business carried out

Expression quality

Normative

Definition: whether the data format and presentation method conform to the unified field specification standard. If they conform to the unified standard, they are qualified, otherwise they are unqualified.

The normalization degree f_norm of data is usually expressed as a percentage, and its calculation formula is as follows: $\begin{matrix} f_{i - j} = \\ {\begin{matrix} 1, a_{i - j} accord with F_{j} \\ 0, a_{i - j} Non conformity F_{j} \end{matrix} (1 \leq i \leq m, 1 \leq j \leq n) \end{matrix}$ (4)

$f_{i} = \frac{\sum_{j = 1}^{n} f_{i - j}}{n} \times 100$ (5)

$normalization (%) = \frac{\sum_{i = 1}^{m} f_{i}}{m}$ (6) where f_i-j represents the normative index value of the j-th attribute value in the i-th row; f_i indicates the degree of data standardization of the i-th row; a_i-j represents the j-th attribute value of the i-th row; F_j represents the reference standard for the j-th data representation attribute; m represents the number of attributes of the data; n represents the sample size of each type of data.

Property integrity

Definition: whether property information of the described object is missing:

$c_{i - j} = {\begin{matrix} 1, a_{i, j} \neq null \\ 0, a_{i, j} = null \end{matrix} (1 \leq i \leq m, 1 \leq j \leq n)$ (7)

$c_{i} = \frac{\sum_{j = 1}^{n} c_{i, j}}{n} \times 100$ (8)

$AI (%) = \frac{\sum_{i = 1}^{m} c_{i}}{m}$ (9) where c_i-j represents the integrity indicator value of the j-th attribute value in line i; c_i is the data integrity rate of line i; a_i,j represents the j-th attribute value of line i; m represents the number of attributes of data; n represents the sample size of each type of data.

When evaluating the accuracy of data for ceramic 3D printing samples, you can select any of the above indicators to evaluate the quality of data for ceramic 3D printing samples.

2.3 Establishing the weight matrix of evaluation index for 3D printing sample data of ceramics

In this paper, the fuzzy algorithm is used to determine the weight of data for ceramic 3D printing samples. In the 1970 s, American operational research scientists put forward a decision-making analysis method that combines qualitative and quantitative analysis, i.e. fuzzy algorithm [16]. It quantifies and models the decision-making thinking process of managers for complex events. The feature of fuzzy algorithm is to divide the complex problems into several levels and factors, and form the multi-level analysis structure model. Through the comparison and calculation of each factor in a certain level, the weight of different factors can be obtained, which provides the basis for data for ceramic 3D printing samples to evaluate the selection of the optimal scheme.

When using the fuzzy algorithm to establish the weight matrix of data for ceramic 3D printing samples, it is usually divided into the following four steps:

–Establish the ladder type fuzzy evaluation index system

For the more complex problem, based on the main factors affecting the quality evaluation of data for ceramic 3D printing samples, the problem is divided into several simple problems. The complex target is divided into several sub element indicators. If the sub element indexes are still complex, the fuzzy decomposition is continued until the element indexes of the last layer are all simple indexes. Thus, a fuzzy index evaluation system is established [2 –4]. The process of establishing evaluation index system is actually the process of decomposing complex problems. In this paper, the target layer of the quality evaluation system of micro data for ceramic 3D printing samples is the quality level of micro data for ceramic 3D printing samples. Due to the complexity of data for ceramic 3D printing samples quality evaluation, the three-tier structure index evaluation system is established as shown in Fig. 1.

–Construct the judgment matrix of pairwise comparison

After establishing the ladder type fuzzy evaluation index system, the judgment matrix is constructed based on the subordinate relationship between the upper and lower layers [12]. That is to say, the element index of the above layer is the fixed criterion, and the relative importance of the factor index of the next layer to the layer is compared by two pairs, and the quantity scale is given. Generally, 1-9 scale method is used to scale the judgment matrix, as shown in Table 3.

Table 3
1–9 scale method

Scale Definition and description Remarks

1 Two elements are equally important to an attribute Nothing

3 I is slightly more important than j Nothing

5 I is more important than j Nothing

7 I is much more important than j Nothing

9 I is more important than j Nothing

2,4,6,8 Intermediate value of the two adjacent standards Nothing

$\frac{1}{b_{n}}$ Inverse comparison of two elements Nothing

Scale	Definition and description	Remarks
1	Two elements are equally important to an attribute	Nothing
3	I is slightly more important than j	Nothing
5	I is more important than j	Nothing
7	I is much more important than j	Nothing
9	I is more important than j	Nothing
2,4,6,8	Intermediate value of the two adjacent standards	Nothing
$\frac{1}{b_{n}}$	Inverse comparison of two elements	Nothing

According to the 1–9 scale method, the relative importance of n elements in the same layer can be obtained, and the judgment matrix can be established.

$I = [\begin{matrix} w_{11} & L & w_{1 n} \\ M & O & M \\ w_{m 1} & L & w_{mn} \end{matrix}]$ (10)

–Calculate the weight coefficient

Based on the above judgment matrix, the maximum eigenvector, which is the weight of each factor, is obtained by normalization, and w = [w₁, w₂, L, w_n]. For the multi-level evaluation system, the weight of each factor index to the previous factor index can be determined from top to bottom, and finally the weight of each factor relative to the target layer can be obtained.

–Check the consistency of judgment matrix

Since the judgment matrix is based on the experience of different individuals, it is inevitable that there are different opinions and judgment errors, so it is necessary to check the consistency of the judgment matrix to ensure the rationality of the judgment matrix [14]. Generally, consistency test includes consistency index, evaluation random consistency index and random consistency ratio.

* The consistency indicators

$C . I . = \frac{λ_{max} - n}{n - 1}$ (11)

where λ_max represents the maximum feature root of the feature matrix; n indicates the number of indicators.

The maximum feature root of the feature matrix is defined as follows:

$λ_{max} = \sum_{i = 1}^{n} \frac{{(AW)}_{i}}{{nW}_{i}}$ (12)

where A represents the judgment matrix; W indicates the indicator weight.

The greater the value of the consistency index is, the closer the judgment matrix is to the complete consistency; the greater the value of the consistency index is, the greater the deviation of the judgment matrix from the complete consistency is. Generally, with the increase of the order of the judgment matrix, the value of the deviation from the completely consistent index caused by human is larger.

* Random consistency index:

For multi order judgment matrix, the average random consistency index is introduced.

* Random consistency ratio:

$C . R . = \frac{C . I .}{R . I .}$ (13)

The ratio between the consistency index C . I . of judgment matrix and the random consistency index R . I . of the same order is called random consistency ratio. When C . R . is less than 0.10, the judgment matrix has acceptable consistency; otherwise, the judgment matrix needs to be adjusted and modified until it has satisfactory consistency.

2.4 Analyze the effective dimension of 3D printing sample data of ceramics

There are many dimensions that affect data validity. Assuming the validity rating dimension is f, each dimension is defined as f₁, f₂, f₃, ⋯ , f_p, and meets the following requirements:

$⋃_{i = 1}^{P} f_{i} = F, f_{i} \cap f_{j} = φ, 0 < i \neq j \leq p$ (14)

Under the background of fuzzy algorithm, because of the 3 V characteristics of data for ceramic 3D printing samples, the possibility of data quality problems becomes higher [15]. Due to the scale and high speed of fuzzy algorithm, there are many sources of data for ceramic 3D printing samples, which may produce conflicting data and incompatible data for the description of the same object. There are also many problems in the generation, transmission and processing of data for ceramic 3D printing samples, such as incorrect and incomplete data for ceramic 3D printing samples. These problems are the main factors that affect the effective use of data [11]. Therefore, in order to evaluate the validity of data for ceramic 3D printing samples scientifically and reasonably, the integrity, correctness and compatibility of data for ceramic 3D printing samples are selected as the evaluation dimensions of data validity.

For different applications, each attribute in data for ceramic 3D printing samples has different influence, so the weight of each attribute is different [22]. For a part of data in the data for ceramic 3D printing samples set, suppose that the data has m attributes, and the weight of each attribute is w_j, that is w₁, w₂, w₃, ⋯ , w_m. w_j satisfies the following requirements:

$\sum_{j = 1}^{m} w_{j} = 1$ (15)

At present, there is no definition for the integrity of data for ceramic 3D printing samples. From the specific application, the integrity of data for ceramic 3D printing samples under the background of fuzzy algorithm is defined as follows:

Definition 1: if a data for ceramic 3D printing samples has n attributes, and each attribute keeps its own part, then the data for ceramic 3D printing samples is complete for these n attributes, otherwise it is incomplete.

Definition 2: integrity is used to indicate the integrity of data for ceramic 3D printing samples. For some data for ceramic 3D printing samples, R_i can be used. The attributes in this part of data for ceramic 3D printing samples are represented by R_ij. t [R_ij] is the value of R_ij, and V [R_ij] is the integrity of R_ij. According to different applications, V [R_ij] can have different forms.

2.5 Building fuzzy data set of 3D printing samples of ceramics

In this paper, fuzzy algorithm is used to evaluate the quality of traffic flow data. Through the establishment of index set, evaluation set, weight set and membership function determined by AHP, the comprehensive evaluation of data for ceramic 3D printing samples quality level is realized. Fuzzy algorithm is proposed by American controller. The evaluation method based on fuzzy algorithm theory is called fuzzy comprehensive evaluation method. At present, this method has formed the relatively complete set of bodies [8].

The basic idea of fuzzy evaluation is to quantify the factors with unclear boundary and not easy to quantify. With the help of the membership degree theory in fuzzy mathematics, the qualitative evaluation is transformed into the quantitative evaluation method. This method can quantify things that are not easy to quantify, such as “excellent, good, medium, poor”, “qualified, unqualified” and other similar fuzzy language scores [23]. Fuzzy evaluation method is mainly used to judge the ceramic process flow in the field of intelligent ceramic 3D printing, but it is still blank in the quality evaluation of data for ceramic 3D printing samples.

There are four steps to build the fuzzy set of data for ceramic 3D printing samples:

–Determine the indicator set of evaluation factors

Determine the indicator set of quality factors of data for ceramic 3D printing samples (Q ={ Q₁, Q₂, L, Q_n }), that is, there are n evaluation indicators.

Data evaluation set is a set of evaluation grades composed of evaluation results made by evaluators for the evaluated objects. In this paper, the quality evaluation set of data for ceramic 3D printing samples is V ={ v₁, v₂, v₃, v₄ }.

–Carry out single factor fuzzy evaluation and establish fuzzy relation matrix

From a single factor, the membership degree of evaluation index U to evaluation set V is determined, and then the fuzzy matrix of the whole evaluation system is obtained, as $R = [\begin{matrix} r_{11} & L & r_{1 m} \\ M & O & M \\ r_{n 1} & L & R_{nm} \end{matrix}]$ . Where r_ij is the membership degree of data quality evaluation index u_i to data quality evaluation set v_i. The fuzzy matrix reflects the degree of subordination of each data quality evaluation index to each level of data quality.

–Determine membership function

The key to determine the fuzzy relation matrix is to determine the membership function of each evaluation index. Membership function is used to describe the membership degree of index to evaluation set, which is usually expressed as u_I (X). When u_I (X) = 1, it means that the indicator is totally subordinate to I; when u_I (X) = 0, it means that a part of the indicator is totally not subordinate to I; when 0 < u_I (X) < 1, it means that a part of the indicator is subordinate to I.

There are many methods to construct membership function, including matrix function, normal function, Cauchy function, trigonometric function, trapezoid function and so on. Because the relationship between accuracy, integrity, effectiveness and data for ceramic 3D printing samples is more in line with trapezoid function, and trapezoid function calculation is simple, so trapezoid function is selected to construct membership function in this paper [6]. Generally, the construction method of membership function is based on the characteristics of various factors, such as the distribution law based on a large number of historical data or the widely recognized and conventional division standards in the industry. At present, there is no unified standard document for traffic data quality evaluation or the classification of related indicators. In addition, the selection of each evaluation indicator also has its own emphasis. Therefore, this paper mainly based on the frequency statistics of a large number of historical data to determine the membership function of each evaluation index.

2.6 Realize the data evaluation of 3D printing ceramic samples

In the evaluation of data for ceramic 3D printing samples, the determination of the weight mainly depends on the expert scoring method, which can make full use of the knowledge and experience of experts in the field to evaluate the relative importance of indicators, so as to obtain the weight. The weight obtained by this method has strong authority and practicability. However, data for ceramic 3D printing samples is dynamic, and the contribution of each index to the importance of the evaluation results may change in different devices and periods. Therefore, a reasonable weight should be a combination of subjective and objective factors.

In this paper, the fuzzy algorithm is used to calculate the index weight, and the influence of dynamic data is considered in the original method, and the dynamic linear correction is made to the weight. In this paper, the fuzzy comprehensive evaluation method is improved by using the data difference driving principle and the linear compensation method of weight. In this paper, the improved method is called dynamic fuzzy comprehensive evaluation method [5], which can not only save the suggestions of experts in the field, but also grasp the dynamic changes of data and the impact of data on indicators in time. When the data is abnormal, it can reduce the impact on the evaluation results; when the data is correct, it can obtain the dynamic development of the traffic state through reasonable.

The evaluation method of data for ceramic 3D printing samples based on fuzzy algorithm is mainly composed of three stages: data preparation stage, comprehensive weight calculation stage and fuzzy decision-making stage, as shown in Fig. 4.

Fig. 4

Evaluation flow chart of data for ceramic 3D printing samples.

In the context of the fuzzy algorithm, data for ceramic 3D printing samples were first collected in the data set for ceramic 3D printing samples. Because the collected data contains a lot of noise information, the fuzzy algorithm was used to preprocess the data for ceramic 3D printing samples. According to the preprocessing principle of ceramic 3D printing samples data, the specific content of the evaluation index of data for ceramic 3D printing samples was determined, and the evaluation system of data for ceramic 3D printing samples was established. The weight of the quality evaluation index of the data for ceramic 3D printing samples is determined by using the fuzzy algorithm. According to the weight matrix establishment step of the evaluation index of the data for ceramic 3D printing samples, the evaluation index weight matrix is established. Because there are many sources of data for ceramic 3D printing samples, there may be data that conflicts with the description of the same object. Under the background of fuzzy algorithm, the validity dimension of data for ceramic 3D printing samples was established. In order to prevent data for ceramic 3D printing samples from being lost during the evaluation process, a fuzzy set of data for ceramic 3D printing samples was built according to the membership function of each evaluation index. Finally, the data evaluation process is designed to realize the evaluation of data for ceramic 3D printing samples based on fuzzy algorithm [13].

3 Results

3.1 Evaluation index weight

Since different ceramic 3D printing sample data evaluation experiments have different indexes, different weights were given to the ceramic 3D printing sample data evaluation indexes, and the evaluation index weights were selected. The selection criteria of evaluation index weights are shown in Table 4.

Table 4
Selection criteria of evaluation index weights

Weight W1 W2

Nozzle diameter 0.1 1.1

Printing speed 0.2 1.0

scanning speed 0.3 0.9

Screw speed 0.4 0.8

Ink volume 0.5 0.7

Print quality 0.6 0.6

Ink volume 0.7 0.5

Wear resistance 0.8 0.4

Screw angular velocity 0.9 0.3

Drying speed 1.0 0.2

Waterproofing property 1.1 0.1

Weight	W1	W2
Nozzle diameter	0.1	1.1
Printing speed	0.2	1.0
scanning speed	0.3	0.9
Screw speed	0.4	0.8
Ink volume	0.5	0.7
Print quality	0.6	0.6
Ink volume	0.7	0.5
Wear resistance	0.8	0.4
Screw angular velocity	0.9	0.3
Drying speed	1.0	0.2
Waterproofing property	1.1	0.1

3.2 Experimental process

Figure 5 shows the experimental process of ceramic 3D printing sample data evaluation.

According to the experimental process, the specific implementation steps of the experiment are determined.

Step 1: set up the experimental environment and get the sample data. Matlab 7.0 software and window 10 operating system were used in the experiment. The experimental data comes from a large-scale ceramic 3D printing product manufacturer, whose product parameters and background data are taken as the experimental sample data.

Step 2: determine the type of data for ceramic 3D printing samples evaluation experiment. According to the evaluation requirements of different types of data for ceramic 3D printing samples, the experimental model is made. The experimental model is saved as a source file format, which is convenient for editing the evaluation results of data for ceramic 3D printing samples.

Step 3: before making the model, each type of data for ceramic 3D printing samples evaluation scheme is divided reasonably to meet the overall needs of the experiment.

Step 4: establish a fuzzy algorithm evaluation platform. Edit the running script of the evaluation system in the computer, associate the experimental equipment with the model, and ensure the normal operation of the whole experiment.

Step 5: Evaluation indexes from ceramic 3D printing sample data were imported into the model, and the weight values of different evaluation indexes were used to calculate the scores and evaluation accuracy of each evaluation index.

3.3 Analysis of experimental results

Using the above experimental process, statistical tables of user ratings and expert ratings of the two evaluation methods were obtained, as shown in Table 5.

Table 5
Statistics of experimental results

Evaluation method of 3D Printing sample data of traditional ceramics Evaluation method of ceramic 3D printing sample data based on fuzzy algorithm

User rating Expert rating User rating Expert rating

Nozzle diameter 54 48 92 79

Printing speed 56 51 94 82

scanning speed 60 52 90 81

Screw speed 54 49 95 89

Ink volume 59 53 96 86

Print quality 62 57 91 83

Ink volume 53 48 94 85

Wear resistance 57 50 93 84

Screw angular velocity 62 54 92 80

Drying speed 54 49 94 90

Waterproofing property 53 47 97 87

	Evaluation method of 3D Printing sample data of traditional ceramics	Evaluation method of ceramic 3D printing sample data based on fuzzy algorithm
Nozzle diameter	54	48	92	79
Printing speed	56	51	94	82
scanning speed	60	52	90	81
Screw speed	54	49	95	89
Ink volume	59	53	96	86
Print quality	62	57	91	83
Ink volume	53	48	94	85
Wear resistance	57	50	93	84
Screw angular velocity	62	54	92	80
Drying speed	54	49	94	90
Waterproofing property	53	47	97	87

From the statistical table of experimental results, it can be seen that when using the traditional evaluation method of data for ceramic 3D printing samples, users and experts have relatively low scores on data for ceramic 3D printing samples, and the expert scores are lower than users’ scores. The average score of traditional data evaluation method for ceramic 3D printing samples is 53.7 points, while when using data evaluation method for ceramic 3D printing samples based on fuzzy algorithm, users and experts have scored more than 80 points for interior design scheme. The evaluation method of data for ceramic 3D printing samples based on fuzzy algorithm gets 88.8 points. Therefore, it can be concluded that the evaluation method of data for ceramic 3D printing samples based on fuzzy algorithm has a high evaluation score for data for ceramic 3D printing samples.

On the basis of the above experiments, a comparison experiment of evaluation accuracy was conducted, and the experimental results are shown in Fig. 5.

Fig. 5

Comparison of evaluation accuracy of different methods.

According to the above figure, the evaluation accuracy of the traditional ceramic 3D printing sample data evaluation method is between 73% and 79%, while the evaluation accuracy of the ceramic 3D printing sample data evaluation method based on the fuzzy algorithm is above 96%, far higher than the traditional method, indicating that the method has good evaluation accuracy, reliability and practicability.

4 Conclusions

In order to solve the problem of low evaluation score and accuracy of traditional methods, a ceramic 3D printing sample data evaluation method based on fuzzy algorithm was proposed. Experimental results show that the method on the users and experts in the indoor design scored in more than 80 points, the evaluation accuracy is above 96%, has the high accuracy evaluation scores and evaluation, which fully shows the fuzzy algorithm is applied to the ceramic 3D printing sample data evaluation has the very good effect, for the further development of ceramic 3D printing technology to provide technical support, can also provide new way for the design of the sample data evaluation methods, to promote further development in the field of 3D printing ceramic, but 3D printing ceramic sample data evaluation method still has a lot of space, In the future, with the development of science and technology, further research on this issue is needed to maximize the evaluation effect of ceramic 3D printing sample data.

References

Cynthia

, Randles

and Sliva

M.D.

et al., The MERRA-2 aerosol reanalysis, 1980 – onward, part i: system description and data assimilation evaluation, Journal of Climate30(17) (2017), 6823.

Yang

, Yang

and Liu

et al., Research on 3D positioning of handheld terminal based on particle swarm optimization, Journal of Internet Technology20(2) (2019), 563–572.

, Shen

and Mi

W.J.

et al., Ship identification algorithm based on 3D point cloud for automated ship loaders, Journal of Coastal Research (2015), 28–34.

Asche

C.V.

, Seal

and Kahler

K.H.

, Evaluation of healthcare interventions and big data: review of associated data issues, Pharmacoeconomics35(7) (2017), 1–7.

Mitra

, Fatakdawala

and Nguyentruong

, Detection of pentosidine cross-links in cell-secreted decellularized matrices using time resolved fluorescence spectroscopy, Acs Biomater Sci Eng3(9) (2018), 1944–1954.

Parente

, Colannino

and Picconi

, Safety of oral alpha-lipoic acid treatment in pregnant women: a retrospective observational study, Eur Rev Med Pharmacol Sci21(18) (2017), 4219–4227.

Yang

, Jiang

and Ang

, A winner-take-all evaluation in data envelopment analysis, Annals of Operations Research278(2) (2019), 1–18.

Wimmer

, Tröscher

A.R.

and Brunner

, Systematic evaluation of RNA quality, microarray data reliability and pathway analysis in fresh, fresh frozen and formalin-fixed paraffin-embedded tissue samples, Sci Rep8(1) (2018), 6351.

Cuevas-Martinez

J.C.

and Yuste-Delgado

A.J.

, and a. trivino-cabrera, cluster head enhanced election type-2 fuzzy algorithm for wireless sensor Networks, IEEE Communications Letters (99) (2017), 1–1.

10.

Sinning

J.M.

, Trying to square the circle? integration of computed tomography data for the evaluation of mitral regurgitation, European Heart Journal38(28) (2017), 2227.

11.

Yang

and Hu

, Evaluation of retrieval methods of daytime convective boundary layer height based on lidar data, Journal of Geophysical Research-Atmospheres122(8) (2017), 4578–4593.

12.

Kaushal

and Khehra

B.S.

, BBBCO and fuzzy entropy based modified background subtraction algorithm for object detection in videos, Applied Intelligence47(4) (2017), 1–14.

13.

Richard

O.A.

, Jakhu

and Zomer

, Adherence and persistence among statin users aged 65 years and over: a systematic review and meta-analysis, J Gerontol A Biol Sci Med Sci73(6) (2017), 813–819.

14.

Gupta

, Mehlawat

M.K.

and Mahajan

, Data envelopment analysis based multi-objective optimization model for evaluation and selection of software components under optimal redundancy, Annals of Operations Research (1) (2018), 1–24.

15.

Kovalev

, Puchkov

N.V.

and Kovalev

S.S.

, The first quantitative evaluation data on vendian metamorphism in the eastern part of the bashkir meganticlinorium, Doklady Earth Sciences483(1) (2018), 1418–1422.

16.

Sevimoglu

, Turanli

and Bereketoglu

, Systems biomarkers in psoriasis: integrative evaluation of computational and experimental data at transcript and protein levels, Gene647 (2018), 157–163.

17.

Lokesha

, Shruti

and Deepika

, Reckoning of the Dissimilar Topological indices of Human Liver, Applied Mathematics & Nonlinear Sciences3(1) (2018), 265–276.

18.

Maximilian

and Hildegard

, Predicting the separation of time scales in a heteroclinic Network, Sciences4(1) (2019), 279–288.

19.

Dariusz

, Comparison of fractional order derivatives computational accuracy-right hand vs left hand definition, Sciences2(1) (2017), 237–248.

20.

Yan

and Li

F.J.

, Noether’s theorems of variable mass systems on time scales, Applied Mathematics & Nonlinear Sciences3(1) (2018), 229–240.

21.

Zou

and Liu

J.A.

, Mutual information-based two-phase memetic algorithm for large-scale fuzzy cognitive map learning, IEEE Transactions on Fuzzy Systems (99) (2017), 1–1.

22.

Cheng

and Fu

, Dielectric properties for differentiating normal and malignant thyroid tissues, Med Sci Monit24 (2018), 1276–1281.

23.

Xie

, Tu

and Xu

, Bacterial distributions and prognosis of bloodstream infections in patients with liver cirrhosis, Sci Rep7(1) (2017), 11482.

24.

Tan

Y.C.

and Zhang

et al., Service performance evaluation using data envelopment analysis and balance scorecard approach: an application to automotive industry, Annals of Operations Research248(1–2) (2017), 1–22.

	Evaluation method of 3D Printing sample data of traditional ceramics		Evaluation method of ceramic 3D printing sample data based on fuzzy algorithm
	User rating	Expert rating	User rating	Expert rating
Nozzle diameter	54	48	92	79
Printing speed	56	51	94	82
scanning speed	60	52	90	81
Screw speed	54	49	95	89
Ink volume	59	53	96	86
Print quality	62	57	91	83
Ink volume	53	48	94	85
Wear resistance	57	50	93	84
Screw angular velocity	62	54	92	80
Drying speed	54	49	94	90
Waterproofing property	53	47	97	87