An Experimental Investigation of Al2O3-40% TiO2 Powder Amalgamated via Atmospheric Plasma Spray Coating onto SS316 Substrate and Parameter Optimization Using TLBO Algorithm

SS316 is a commercial stainless steel. MTBF (Mean Time Between Failure) of SS 316 wear prone areas can be effectively increased by ceramic coating. The coating thickness, surface roughness, coating microhardness, abrasion rate, and coating porosity decides the quality and durability in ceramic coating. The current research work explains an experimental investigation to optimize the Atmosphere Plasma Spray process input parameters of Al2O3-40%TiO2 ceramic coatings. Threelevel L18 Orthogonal Array (OA) design of Experiments (DoE) is used to conduct the current work. The main input parameters considered in the current study are nozzle distance, substrate speed, arc current, carrier gas flow, and coating powder flow rate. The output parameters considered are coating thickness, surface roughness, coating microhardness, abrasion rate, and percentage of porosity. Mathematical models are generated for individual output parameters. AHP (Analytical Hierarchy Process) is effectively used to find out weights for individual output parameters treating them as objective functions, and a combined objective function is generated.


Introduction
To achieve increased reliability and performance of damage prone industrialrelated components, surface engineering is hugely now applied using large field of new technologies. The quest for higher efficiency and productivity across the entire spectrum of manufacturing and engineering industries has ensured that most of the machine components are subjected to highly harsh environments during routine operation [1,2]. The high deterioration of parts and their ultimate failure has been traced to material damage bought in by hostile environments like high relative motion between mating services, corrosive media, extreme temperatures, and cyclic stresses. As a result of the above, the concept of applying engineered surfaces capable of combating the high degradation phenomena like wear, corrosion, and fatigue to improve component performance, reliability, and life cycle has gained high acceptance in last one decade. To act as a last line of defense, a proactive coating deposited to act as a perfect barrier between the initial surface of the component and the aggressive environment that is exposed during routine operation is now globally acknowledged as an attractive and effective solution to significantly reduce damage. 316 stainless steel is widely used in chemical/petrochemical industry, food processing, pharmaceutical equipment manufacturing, medical devices fabrication, in potable water/wastewater treatment/marine applications and architectures. SS316 composition consists of Chromium 10-14%, Nickel 2-3%, carbon 16-18%, and Molybdenum. In general, the addition of about 2% Molybdenum in SS316 stainless steel provides excellent level of resistance against pitting corrosion and stress corrosion cracking especially in a saline environment compared to SS304. Abrasion and wear resistance is the type of damage which needs to be improved during its application in industrial environment. Abrasion resistance capabilities of SS316 can be highly enhanced by providing a coating of Al 2 O 3 -40%TiO 2 on an SS316 steel surface by means of atmospheric plasma spraying.
Hence Al 2 O 3 -40%TiO 2 coating on SS316 surface makes it an ideal combination to combat today's highly drastic and hostile industrial environments where assets are sweating to achieve target of productivity and yield which is always more than 100% of the rated capacity. Even though there is a high level of interest and expectations regarding the fundamentals and development of atmospheric plasma arc spraying, there is a huge gap of reliable as well as dependable models that correlates final engineering properties of coatings such as surface roughness, microhardness, porosity, etc. with variations in critical input process parameters and geometry of deposition process [3].
In the present work, Al 2 O 3 -40% TiO 2 is coated on SS316 substrates and all the desired critical output parameters are measured and recorded for further analysis.

Experimentation details
Justification for the usage of Al 2 O 3 -40% TiO 2 amalgamated powder. Al 2 O 3 -40% TiO 2 amalgamated powder is used for all the experiments conducted during this research work. The use of Al 2 O 3 -40% TiO 2 justified as it helps to decrease the melting temperature of amalgamated composite powder and hence the final coating exhibits very low porosity % and enhanced fracture toughness compared to 97/3 and 87/13 Al 2 O 3 and TiO 2 combinations. The coating generated with Al 2 O 3 -40% TiO 2 amalgamated powder also possesses high dielectric strength, enhanced wear and heat resistance [4].

Development of experimental plan
As per the literature review, the following process parameters play a deciding role in the Al 2 O 3 -TiO 2 atmospheric plasma spray process on various substrates.
Major input parameters of a plasma arc spray are below: After considering the critical requirements of the industry, review of facilities available and thorough literature survey, the following input parameters with three levels are considered for experiments as shown as Table 1.
Similarly, the output parameters selected are: coating thickness, surface roughness, microhardness, abrasion rate, and porosity %.

Design of experiments
Orthogonal array experimental design proposed by Taguchi can be efficiently used to examine the effect of different input parameters on the critical performance characteristics linked output parameters through compact set of experiments [5]. In-depth understanding of the process, including the minimum, maximum, and current value of the parameter is required to decide the set of input parameters that are highly affecting a process as well as the levels at which these parameters should be varied [6].
After thoroughly understanding the number of input parameters and the number of levels a proper compact can be finalized using the array selector as shown in Table 2, by looking at the column and row corresponding to the number of parameters and number of levels.
An L18 orthogonal array is used to carry out all the experiments as shown in Table 3 considering the five input parameters with three levels. Table 4 shows the complete experiment plan.

Experimental steps
Al 2 O 3 -40%TiO 2 powder from H C Starck, USA is utilized for coating for all the set of experiments. SS316 substrates are prepared with 27.5 mm diameter and 3 mm thickness plates and pieces of size 75 Â 25 Â12 mm are prepared for coating so as to conduct the Abrasion test [7].
Each experiment is carried out with three substrate samples. The substrate samples are assembled in one specially fabricated cartridge [8]. To avoid nonuniformity in thickness, the substrate samples are ground to achieve relatively good finish. Later, sand blasting is carried out on the surface of all substrate samples to ensure proper removal of oxides and other impurities [9,10]. For sand blasting, fused alumina of grit size 60 μm from Carborandum Universal is used [11,12]. Al 2 O 3 -40% TiO 2 powder is deposited on the substrates by using a controlled atmospheric plasma stray system of Metco USA through an SG 100 model Plasma Gun. To ensure the removal of moisture, the amalgamated powder is preheated before the plasma coating process up to 110°C [13].
As pre the L18 orthogonal array DoE, the experiments are conducted. The coating facility arrangement is shown in Figure 1. The substrate samples are cleaned after the coating with ethanol and properly dried to eliminate accumulation of moisture [14]. As per ASTM B499-92014 [15], coating thickness is measured. An ultrasonic thickness gauge is used for this purpose. Using Mitutoyo surface roughness tester SV-C3100, surface roughness is measured as per ASTM D127 2013 [16]. The surface testing probe is applied on the coating surface for a length of 15 mm with a pitch of 0.001 mm. The scanning speed is kept constant at 2.0 mm/s. Microhardness is measured as per ASTM B 578-872,015 [17] by Metatech MVH-1. To measure porosity as well as microhardness, the samples are sectioned by wire cutting as well as slow speed grinding.
Later, molds are prepared using Bain mount-3 molding machine supplied by Chennai Matco. The substrate surfaces on the mold are polished with emery papers ranging from 150 grit up to 2000 grit so as to get scale free and super finished surface. The porosity is measured using an Image Analysis Software as shown in Figure 2. The JPG images of the coatings are taken by using an electronic microscope OLIMBUS GX51 as per ASTM E2109-012014 [18]. A sample image of the sectioned coating surface is shown as Figure 3. Abrasion rate is measured on Abrasion test rig TR-50 as per ASTM G65 2000 [19] with 1000 g. as pre-load. Figure 4 shows one of the abrasion test in progress. This test method covers laboratory procedures for determining the resistance of coating materials to scratching abrasion by means of the dry sand/rubber wheel test. Abrasion test results are reported as weight loss in grams for 20 revolutions of the wheel for a particular substrate. Materials of higher abrasion rate and high weight loss show a lower abrasion resistance.

Details of coating output parameters
The measured values of coating thickness, surface roughness, microhardness, abrasion rate, and porosity % are given in Table 5.
The values of roughness, abrasion rate and porosity are considered nonbeneficial and the values of thickness and hardness are considered as beneficial.

Confirmation experiments
For conducting confirmation tests, three trial samples of SS316 are used. The random values for all the input parameters, in between the maximum and minimum levels are taken to conduct the confirmation tests. The predicted values using the proposed model along with the measured output parameters values for three samples are given in Tables 6-10. The percentage variation between the actual and predicted values are also shown in the tables.

SN analysis
To determine the effect each variable has on the output, the signal-to-noise ratio, or the SN number, needs to be calculated for each experiment conducted. In the equations below, y i is the mean value and s i is the variance. y i is the value of the performance characteristic for a given experiment. More details about SN analysis is given in https://google site sn analysis design of experiments.
SN ratio values for coating thickness on SS316 substrate are calculated for each parameter and level. The values are tabulated as shown in Table 11.
Spray distance has a significant effect on the coating thickness. Carrier gas flow is the next dominant parameter in the case of coating thickness. Similarly, SN ratio values are calculated for surface roughness for each parameter and level for all the output parameters as shown in gas flow has a significant effect. Substrate rpm is the next influencing parameter as per R value. The spray distance has a significant effect on the microhardness as shown in Table 13. The next dominant parameter in the case of microhardness is rpm. For coating abrasion rate, spray distance is the most dominating parameter and then comes the rpm as shown in     shown in Table 15, the most dominating parameter is carrier gas flow. The next parameter which effects the most is powder flow rate.

Application of teaching learning based optimization (TLBO)
An advanced optimization method, known as Teaching-Learning-Based Optimization (TLBO) is applied, to determine the best values of input parameters to obtain global optimum output parameters. TLBO is applied individually to each of the developed mathematical models given by Eqs. (1)-(5). Rao et al., proposed TLBO, which is based on the effect of influence of a teacher on the output of learners in a class. Teaching-learning ability of teacher and learners in a class room is mimicked in this algorithm [20,21]. There are two modes of learning in this algorithm, interacting with other learners (known as learner phase) and through teacher (known as teacher phase).One of the attractive features of this algorithm is its algorithm-specific parameter-less concept. The algorithm is widely preferred among researchers due to its simplicity and its ability to provide the global optimum solutions in comparatively less number of function evaluations. Further details about the TLBO algorithm can be found at https://sites.google.com/site/tlborao.
To execute TLBO algorithm for the optimisation of individual objective functions, a population size of 10 and 100 number of iterations with 30 independent runs is considered. The global optimum values for individual objective functions of T (Coating thickness), R (Surface roughness), H (Microhardness), Ab (Abrasion rate) and Po (Porosity %) obtained after applying TLBO are given in Table 16.

Conclusion
In the field of atmospheric plasma coating with Al 2 O 3 -40%TiO 2 , mathematical modeling and its optimization is very rare. Mathematical models are generated in the present work using regression analysis for all the output parameters in terms of input parameters. The mathematical models developed will work as effective tools for manufactures to predict the effect of input parameters on output parameters within the considered ranges. Based on this model, they can take decisions and hence costly trials can be avoided to a very large extent. Confirmation tests are also carried out in the present work for each of the output parameters. The confirmation tests have given near about the same values compared to the predicted values and the percentage of error is negligible.
The optimization is effectively carried out using teaching-learning-based optimization (TLBO) algorithm for each output parameter individually. A combined objective function is generated and this combined objective function can be again optimized using TLBO algorithm to get global optimum values of input parameters considering all the output parameters simultaneously. The TLBO algorithm has proved its effectiveness and simplicity in solving the multi-objective optimization problems. The AHP method is applied to decide the weights for the individual objective functions in the combined objective function in a systematic way and it takes into account the preferences of the decision maker. It is also concluded that a change in the weights of the individual objective functions in the combined objective function may give different sets of optimum values of input parameters.

Author details
Thankam Sreekumar Rajesh SVNIT, Surat, Gujarat, India *Address all correspondence to: rajeshtsreekumar@gmail.com © 2020 The Author(s). Licensee IntechOpen. This chapter is distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/ by/3.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.