عارف مردانی کرانی

In this research, field experiments were carried out in two types of soil (sandy clay loam and clay) to model the traction performance of the tractor considering drawbar power, rolling resistance, and traction efficiency, using deep learning and convolutional neural network while having some parameters such as soil type and conditions, tool parameters, and operation parameters. The tests were conducted within each soil texture in the form of factorial tests based on the randomized complete block design (RCBD) in triplicates. The tests were done in various moisture levels (8-17% for dry soils and 18-40% for moist soils), tractor forward speed (1.2, 1.6, 1.8, and 2.2 km h-1), working depth (30 and 50 cm), the number of pass (2 and 6 times), and tire inflation pressure (20 and 25 psi). The cone index, dynamic load, draft force, and moisture content were measured in each tests. The networks designed to model the drawbar power, rolling resistance, and traction efficiency were of convolutional neural network type. Various algorithms such as Sgdm, Adam, and Rmsprop were utilized to train the network. The results showed that the neural network developed by Sgdm algorithm outperformed the others. Therefore, this algorithm was utilized for the modeling process. Statistical criteria such as R2 and MSE were also employed to evaluate the performance of the network. For the drawbar power, the 8-499-499-1 architecture showed the best performance with R2=0.9953 and MSE=0.0016. Concerning the rolling resistance, the best performance was observed in 8-301-305-1 architecture with R2=0.9903 and MSE=0.0039. The best performance for the traction efficiency was obtained by 8-371-371-1 architecture with R2=0.9888 and MSE=0.003. The results showed that these networks can be used to model parameters by removing convolution layers and reducing dimensions.

In this study, the possibility of using solar energy for recycling food waste in the West Azarbaijan region was investigated. The method of waste distillation was assumed as the basis for the work, and a system for distillation of wastewater was developed using heating from a solar collector. In the design of the machine, another source of energy was not required to flow into the system. The machine was equipped with an electric element for water heating to perform preliminary tests and performance evaluation. Tests were made to estimate the capacity of the machine in four levels of input wastewater temperature (representing four levels of ambient temperature in different seasons) as well as three levels of thermal evaporation of water. The results indicate increased distillation performance by increasing the initial temperature of the wastewater and also increasing the thermal power of the device. By matching the radiation conditions in the studied area, based on the weather data and test results of the device with the electric element, the distillation capacity of the machine was estimated for the months of the year. According to the research, the maximum distillation capacity of the machine is projected to be about 50 liters in July. The annual capacity of the device is estimated at around 10,000 liters, which is remarkable considering the importance of environmental wastewater treatment as well as the economic benefits of using the free energy of the sun.

Due to the numerous variables that may influence the soil-machine interaction systems, predicting the mechanical response of soil interacting with off-road traction equipment is challenging. In this study, deep neural networks (DNNs) are chosen as a potential solution for explaining the varying soil sinkage rates because of their ability to model complex, multivariate, and dynamic systems. Plate sinkage tests were carried out using a Bevameter in a fixed-type soil bin with a 24 m length, 2 m width, and 1 m depth. Experimental tests were conducted at three sinkage rates for two plate sizes, with a soil water content of 10%. The provided empirical data on the soil pressure-sinkage relationship served as the basis for an algorithm capable of discerning the soil-machine interaction. From the iterative process, it was determined that a DNN, specifically a feed-forward back-propagation DNN with three hidden layers, is the optimal choice. The optimized DNN architecture is structured as 3-8-15-10-1, as determined by the Grey Wolf Optimization algorithm. While the Bekker equation had traditionally been employed as a widely accepted method for predicting soil pressure-sinkage behavior, it typically disregarded the influence of sinkage velocity of the soil. However, the findings revealed the significant impact of sinkage velocity on the parameters governing the soil deformation response. The trained DNN successfully incorporated the sinkage velocity into its structure and provided accurate results with an MSE value of 0.0871.

Seedbed preparation, seeding, and transplanting are usually based on mechanical soil tillage. Tillage by cutting, mixing, overturning, and loosening the soil can modify the physical, mechanical, and biological properties of soil. These days, because of soil protection, the use of tillage tools is less and less recommended, and some implements such as cultivators are preferred to primary tillage tools such as plows. Experimental study of soil-tool interaction and field measurements of the mechanics of tillage tools are usually time-consuming and costly. On the other hand, the variety of variables and uncontrolled conditions add other dimensions to the complexity of this method. Also, the experimental and analytical methods do not have a comprehensive view of stress distribution and soil deformation in the soil-tool interaction process.

The main purpose of this study is to validate the results of numerical simulations in two phases of experimental tests: in soil bin environment and in finite element computer simulations. Experimental tests were performed in the soil bin environment of the Department of Mechanical Engineering of Biosystems, Urmia University, which has a soil bin facility with dimensions of length and width of 24 and 2 m, respectively, and has clay loam soil. Before experimental tests, soil preparation was performed by using some special tillage implements (harrow, leveler, and roller) which were attached to the soil bin (Figure.1). For experimental tests, a mechanism set consisting of two cultivator blades with a width of 15cm, a length of 20cm, and at a spacing of 35cm from each other was prepared and constructed. The relevant mechanism is designed to have the ability to change the tillage depth. Data were collected at three different soil depth levels of 6, 10, and 14cm in the soil bin with three replications. Data recording was performed using a 10-channel data logger with load cell connectivity and data storage ability. Also, in this study, the Drucker-Prager model as a finite element simulation method was used to calculate the stress during the soil-tool relationship. ABAQUS 6.10.1 software was used to simulate the cultivator tine. To solve the problem, the soil parameters were defined as presented in Table 1, and then the interaction between the soil-tool model and the necessary constraints, including boundary conditions, were defined. In the next step, meshing was applied to the constructed model.

In the results section, first, the results related to the amount of traction force required for the tillage tine in the simulation were calculated and then compared with the soil bin experimental tests. The traction force of the finite element simulation results for three tillage depths of 6, 10, and 14 cm in three principal directions is shown in Figure 4. A comparison of simulation and experimental results showed that there is a good agreement between them. In comparison, the simulation error range of the three depths of 6, 10, and 14 cm has shown 7.3, 5.6, and 4.16% at a speed of 2.5 kmh-1, respectively, as the velocity studied in this research. In the next section, the results of stress distribution contours in the soil and finally the overlap of the blade effect were discussed. Figure 6 shows the status of stress contours at three depths. By increasing the depth of the tine at the three depth levels studied, the stress range is shifted from the soil surface to its depth. For this purpose, at the maximum depth studied in this study (14 cm), it shows that the stress propagation to the soil surface is less than at other depths. Also, with decreasing depth, for a depth of 6 cm, the maximum stress was on the top soil surface, in other words, more deformation was seen on the soil surface.

Comparing the simulation results for predicting traction force with the results of experimental tests has led to relatively acceptable results and the maximum traction force prediction error at different depths has been about 7.3%.The distribution of stress in the soil was observed due to the tine depth. The highest intensity of stress propagation was observed at the soil surface; and the highest soil surface deformation at a depth of 6 cm. With increasing depth, both parameters of stress and soil surface deformation have decreased. According to the results of the studied blades, it is better to use these types of tillage tools only at lower depths. Also, in evaluating the overlap of the soil loosening zone in the side-by-side tines, it proves the superiority of the tine performance at lower depths.

Tracks have advantages over wheels, part of which is related to the traction performance of the machine, and the other part is associated with reducing soil compaction. In this study, an experimental comparison of wheel and track in terms of vertical soil stress and vertical deformation of the soil during multi-passes has been made. This study used two-wheel and track testers in a loam-clay soil bin environment. The amount of vertical stress in the soils at 8, 16, and 24 cm depths was measured and recorded using a stress measuring system during wheel (track) multi-passes at five levels 1, 3, 7, 15, and 20. Also, the amount of vertical rut depth was measured during the movement of wheel and track passes. The most significant advantage of the track over the wheel on soil compaction occurred at depths of less than 10 cm, and the amount of vertical stress in the soil was only about 40% of the wheel. At greater depths, the wheel and the track were almost the same creating stress, so that at a depth of more than 24 cm, more than 90%, the vertical stress under the wheel and the track were similar. The vertical deformation of the soil showed a significant difference between the wheel's performance and the track. The amount of soil rut depth due to the passage of the track at all levels of the pass number did not exceed 45% of rut depth due to wheel passage.

Potato is an important vegetable that grows all over the world and is considered as an important product in developing and developed countries for human diet as a source of carbohydrates, proteins, and vitamins. This product is native to South America and its origin is from Peru, and after wheat, rice and corn, it is the fourth product in the food basket of human societies. According to the statistics of the Food and Agriculture Organization of the United Nations, the area under cultivation of this crop in Iran in 2017 was 161 thousand hectares and the crop harvested from this area is about 5.1 million tons. Traditional methods of determining potato varieties were based more on morphological features, but with the production of new products, there was a need for methods that were faster and more recognizable. Meanwhile, high-performance artificial neural network can be used to classify cultivars. Artificial neural network can classify and detect cultivars, is flexible and is used in most agricultural products. Azizi conducted a study on 120 potatoes in 10 different cultivars using a visual and image processing machine with a MATLAB R2012 software toolbox to detect texture, shape parameters and potato cultivars. First, potato cultivars were classified using LDA method, which obtained 66.7% accuracy. This method also erred in distinguishing the two cultivars Agria and Savalan and also classified the two cultivars Fontane and Satina in other classes. They also used artificial neural networks to classify potato cultivars, in which the network was 82.41% accurate with one hidden layer and 100% accurate with two hidden layers. In this study, it was found that different types of potatoes can be identified and identified with a very high level of accuracy using the three color characteristics, textural and morphological features extracted by the visual machine and the use of a non-linear classifier artificial neural network. Categorized.By determining and examining the existing relations between the force and the deformation of agricultural products up to the point of surrender, the range of forces harmful to fruit can be determined so that harvesting and transportation machines are designed in such a way that the forces from them do not exceed this range. On the other hand, one of the ways to determine the degree of ripeness of the fruit is to touch and press it with the thumb, which is an experimental way and depends on the skill of the person touching. The mechanical penetration test of the fruit can be an indicator to check the ripeness of the fruit by quantifying this diagnosis and using this diagnosis to determine the optimal harvest time.

First, 5 different potato cultivars were prepared from Ardabil Agricultural Research Center and kept at a temperature of 4-10 ° C. One day later, 21 samples of each potato cultivar were prepared using a cutting cylinder and then data were collected. To determine the toughness of the samples, the Centam device available in Mohaghegh Ardabili University was used. Each potato cultivar was subjected to compressive force at three levels of loading speed of 10, 40 and 70 mm / min with 7 repetitions. Then the amount of toughness was calculated according to Equation (1). Then linear diagnostic analysis (LDA) and artificial neural networks (ANN) were used to classify potato cultivars. LDA is a supervised method used to find the most distinctive special vectors, maximizing the ratio of variance between class and within the class, and being able to classify two or more groups of samples. ANN and pattern recognition were used to find similarities and differences in the classification of potato cultivars. For this, 1 neuron was considered for the input layer, the hidden layer with the optimal number of neurons will be considered and five output neurons with Depending on the number of output classes the target will be considered. In network training, the logarithmic sigmoid transfer function and Lunberg-Marquardt learning method were used (Figure 4), and the error value was calculated using the mean square error. For learning (70%), testing (15%) and validation (15%) all data were randomly selected. Training data was provided to the network during the training and the network was adjusted according to their error. Validation was used to measure network generalization and completion of training. Data testing had no effect on training and therefore provided an independent measurement of network performance during and after training. All of the calculations and matrix classification were performed using MATLAB R2018a and X 10.4 Unscrambler software.Toughness in 5 different potato cultivars was obtained using Centam machine and Equation 1. The values obtained for the toughness of 5 potato cultivars were analyzed using Mstatc software. The results of analysis of variance were significant for the toughness of 5 different potato cultivars at the level of 1% and its coefficient of variation was 2.28. LDA and ANN methods were used to detect potato cultivars based on the values calculated for toughness. Detection results of cultivars using LDA were equal to 70.48% (Figure 6). Also, the accuracy of ANN method according to the perturbation matrix was equal to 72.4% (Figure 7).

In this study, the amount of toughness for 5 different potato cultivars was calculated using Centam machine available in Mohaghegh Ardabili University with the help of Equation 1. Chemometrics methods including LDA and ANN were used for qualitative and quantitative analysis of data to identify and classify potato cultivars. Thus, LDA and ANN were able to identify and accurately classify different potato cultivars with an accuracy of over 70%. The obtained toughness has the ability to be used as a method to distinguish different potato cultivars. The use of this method in identifying potato cultivars will be very useful for factories such as chips factory and processing units, and it is also expected that similar methods related to mechanical properties such as crispness and stiffness with the help of chemometrics methods to optimize production and The processing of agricultural products should be used in the food industry, which has led to more customer friendliness and, in addition, can reduce agricultural waste.

One way to reduce soil compaction is to add organic matter and to manage the field traffic. In this research, farmyard manure was incorporated into clay soil with rates of 0, 45, 60, 90 ton ha-1. After 6 months (September to March), at different numbers of tyre passes of 1, 6, 11 and 16, and three soil moisture contents of 8%, 11% and 14% soil compaction was evaluated measuring soil bulk density and soil sinkage. Experiments were conducted in the soil bin at the Urmia University under a single trector’s tire 220/65 R 21l under a constant load of 4 kN, inflation pressure of 110 kPa and at a forward velocity of 2.88 km hr-1. It was found that at manure rate of of 90 ton ha-1, comparing to no-manure treatment, soil bulk density and soil sinkage decreased by 14.7 and 6.94 percent, respectively. Also, increasing the number of tyre passes from 1 to 16 and increasing soil moisture content from 8 to 14 percent, increased soil bulk density 7.21% and 7.92%, respectively. For neural network modeling multilayer perceptron network with six neurons in the hidden layer with sigmoid transfer function and linear transfer function for the output neuron was used. Comparison of neural network output and experimental results showed high correletion with correlation coefficient of R= 0.99 between them. The mean square error (MSE) of the model and mean absolute percentage error of the system (MAPE) were 0.0119071 and 0.0009641,respectivly, which showed high accuracy of neural network to model soil compaction.

The main objectives of this study is to determine the dynamic impact force imparted on a lugged tire due to vertical load, travel speed of the tire, shape and height of the obstacles, and slippage of tire. The tests have been carried out by means of a soil bin facility. Three shapes of triangular, trapezoidal and curved obstacles were used in the study each at three heights of 2, 3 and 4 cm while two wheel load levels of 1 and 2 kN were considered. These results show that by increasing vertical pre-load on the tire, the total dynamic impact force is increased. This increase could be attributed to dependence of vertical inertial forces to the dead load of the wheel axle. The result of analysis shows linear relation between the tire travel speed and total dynamic force. At constant obstacle shape and height, as travel speed increases, the impact force imparted to tire increases. The dynamic force generated in vertical and horizontal directions have shown significant differences. In the other words, increasing height of obstacles result in the increase of vertical component of tire velocity and then increasing momentum of tire which led to enhancement of vertical force.