1 Introduction

Refrigerated display cabinets (RDCs) are great and vital as they are part of the cold chain for food products [1]. It is easier to reach the desired weather conditions in CRDCs compared to open RDC (ORDC) [2]. Particularly the temperature and humidity are the conditions that should be provided in the pastry products.

After making the pastry products, it may be necessary to display a certain period of time in the store for sale, if this time is too long and the necessary storage conditions for the product are not provided, the pastry products may deteriorate or loss mass. The ideal conditions are adversely affected due to the not achievable ideal relative humidity range, the temperature fluctuations and the door of cabinet which is often opened for product sales. Therefore, the shelf life of products in CRDC is shortened.

For refrigerated foods, generally between − 1 and \(+\,5\,^{\circ }\hbox {C}\) is the acceptable temperature range [3]. Therefore, the temperature and the humidity of the environment where the cake will be hold must be kept at a certain value. Ideal temperature and humidity can be considered as \(+\,4\,^{\circ }\hbox {C}\) and 80% RH, respectively, in order to prevent growing pathogenic microorganisms. Figure 1 shows how pathogenic bacteria grows according to the temperature change.

Fig. 1
figure 1

Bacterial growth rate for human pathogens at different temperatures [4]

Full size image

During the product display in RDCs, moisture loss or gain will continue to occur until the environment and food components (cream and etc.) that surround the food reach the thermodynamic balance [5]. Due to dehydration, the appearances of unwrapped foods are deteriorated and limit the shelf life. Consumers are preferred newly loaded products compared to the product that have been displayed for some time [6]. Furthermore, the evaporation of water from the product is the mass loss of the salable product. In a survey conducted, attention was focused to the importance of relative humidity. Relative humidity decreased from 95 to 40% caused significant mass loss of food [7]. Not only does the humidification process have a positive effect on the microbial quality of the products, but also extends the shelf life. Using humidification equipment to increase the humidity of the cabinet in which the product is stored is a way of reducing the problems mentioned [8].

Consumers prefer indoor CRDCs because of pastry products in Turkey usually displayed as unpacked. Traditional CRDCs which are produced in Turkey have generally only on–off temperature control. There is no humidity control, and humidity is supplied by fans that are directed to the water at the bottom of CRDC. The humidity value is at most 65% RH, and the ideal humidity value (80% RH) is not reached. The basic motivation of this study is to reach the ideal humidity and temperature values in shorter time than traditional system and to preserve these values. In this way, it is ensured that the shelf life of the cakes is longer than it was before.

Microbiological safety of chilled foods is the main key factor and requires product design and cold chain management. Many technology areas are used to make safe products and ensure shelf life [9]. In recent years, computer and communication technology has seen great progress based on real-time status information on cold stores and RDC. RDCs can have integrated temperature and humidity control to extend the shelf life of non-pre-packaged foods. Thus, the measurement and control of food environment conditions can be part of the safety and quality system [10]. Embedded systems are computer systems designed for a specific purpose. These systems include not only hardware, but also embedded software [11]. Thanks to this embedded software and hardware, CRDC can be controlled depending on the critical values of temperature and humidity.

In this study, an EFLCS that can easily be mounted to compatible current CRDCs without too many changes was designed and implemented. Developed EFLCS simultaneously receives information from different sensors and simultaneously activates different actuators. Contribution to the development of alternative solutions to meet the current requirements for CRDCs such as desired RH and temperature values was aimed.

There are some studies in the literature related to control system for ORDC in various countries that are out of scope of this study [12,13,14,15,16,17,18]. Because, ORDCs do not have doors and use different techniques such as air curtain and air guiding strip, to replace the doors and to provide the air conditioning. The parameters to be controlled are different from the parameters of CRDC. Additionally, there are not many studies conducted with control system for CRDC that are similar to developed EFLCS in this study. Experimental and numerical control studies for CRDC in recent years are shown in Tables 1 and 2, respectively [19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38]. When both tables are examined, it has been seen that the methods used for controlling aims are FLC, mode control, PID control, fuzzy expert system, On–off control and sliding mode control. Additionally, the controlled parameters are mainly temperature and humidity [19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38]. Only several studies [24, 29, 35, 37] have one additional control parameter. At this point, it should be declared that there is no benchmark to compare CRDC studies [19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38] to each other since different sizes of areas (from an incubator to big cold storage) and zones (food storage, health, livestock and etc.) were air-conditioned. It means that CRDC studies are domain-specific ones and there is no way of comparing the time for reaching target values, the stabilities, etc. with previously developed systems or studies. Thus, the authors of these papers [19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38] only compared their methods with traditional systems applied on their areas. Due to these reasons, developed EFLCS was compared with traditional system that exists on CRDC used in this study.

Table 1 Similar experimental studies in recent years
Full size table
Table 2 Similar numerical studies in recent years
Full size table

2 Material and Method

The embedded system can be defined as microcontroller /microprocessor-based system designed to control a function or function range. The embedded system includes: a processor capable of performing the required tasks within the system, a memory that stores the software and the data generated during the operation, peripheral units for communicating with the outside world, a software for efficiently running the system, an algorithm which has the ability to interpret the sensors data and generate the control signal according to this data [39]. In general, an embedded system is implemented by: (1) circuit design and simulation, (2) experimental demonstration of the designed circuit (on a breadboard), (3) application of the circuit (soldering), (4) packaging [40].

Fig. 2
figure 2

General scheme of design

Full size image

Microcontroller needs to be programmed to operate according to certain parameters. FL, which can incorporate human thinking ability and experience into the control system, promises both to fix the system in a shorter time and to provide oscillation at lower intervals. FL is also a popular tool that provides decision support for analyzing data classes and makes it flexible by analyzing ambiguous data [41].

Developed EFLC system is aimed for keeping temperature and humidity inside a CRDC manufactured in Turkey at approximately \(+\, 4\,^{\circ }\hbox {C}\) and 80% RH, respectively. Input variables are temperature and humidity information, whereas output variables are speed levels of fans (3) and the steam level of UA. The temperature is provided by evaporative cooling, and the humidity is supplied by UA as cold steam. The fans that are placed near the evaporator are used to reach the desired level of CRDC temperature easily. The temperature and humidity information is transferred by the SHT11 sensor to the PIC18F4620 that is programmed with FL approach. Symmetric fuzzification, Mamdani inference mechanism and weight average method for defuzzification are used for programming PIC18F4620. Depending on these transferred values, the fans and UA are operated to provide the necessary air conditioning for the cake. The gas compressor is on–off controlled in the range of 3–\(5\,^{\circ }\hbox {C}\) to keep the inside temperature of CRDC at \(4\,^{\circ }\hbox {C}\) by relay G5LE-1-ACD model of OMRON company. In addition, the values of all input and output variables can be monitored instantaneously through the interface screen. The scheme of developed EFLCS is shown in Fig. 2.

2.1 Control Card with Microcontroller

Microchip’s PIC18F4620 40-pin microcontroller unit (MCU) has 64K flash, 1K EEPROM, Capture/Compare /PWM modules and built-in USART support. It also has a 10-bit analog-to-digital converter module. This module has 13 inputs and provides conversion of an analog input signal to a corresponding 10-bit digital number. The PIC18F4620 is operated in three modes. These modes are Normal, Idle and Sleep.

The PIC18F4620 microcontroller is selected to develop EFLCS because it has built-in USART support, many digital outputs, 64 kB program memory and different operating modes.

The input/output ports on the control card enables PIC microcontroller to communicate with peripheral devices and the computer. The data cables from the sensors are connected to the input ports of the control card, the fans and UA and the relay pins are connected to the output ports. The control card also has a COM port. The control card can be connected to the computer via this port. Thus, the interface program on the computer can display and record the input/output data. This function has no effect on the operation of EFLCS. FL-based software on the PIC18F4620 evaluates the information from the sensors and sends the control signals to the fans, relay and UA via this card for the purpose of making the air conditioning of CRDC. Therefore, the control card has the most important function in developed EFLCS.

Fig. 3
figure 3

SHT11 connection circuit [42]

Full size image
Fig. 4
figure 4

a YASKAWA V1000 model inverter/driver and b YASKAWA V1000 control inputs [43]

Full size image

2.2 Sensors

The SHT11 sensor measures temperature and humidity values simultaneously in an environment. For developed EFLCS, eight expansion slots were integrated to control card to make connection with SHT11 sensors. Since the measurements of temperature and humidity in five different regions of CRDC (lower left, lower right, upper left, upper right and near evaporator) were needed, only five of the expansion slots were used to connect five pins of SHT11 sensors. Fifth sensor on the evaporator was used to increase the efficiency of evaporator by minimizing the snow on the evaporator surface. For efficient air conditioning system, the difference between the average temperature of the cooled area and temperature of cooler liquid inside the evaporator must not exceed \(5^{\circ }\hbox {C}\). In case of having the temperature difference more than \(5^{\circ }\hbox {C},\) the water steam starts condensing. This causes snow on the evaporator and decrease in relative humidity ratio of the cooled area. To eliminate this situation and to keep this temperature difference under \(5^{\circ }\hbox {C},\) air circulation must be provided between the evaporator and cooled area. Figure 3 shows SHT11 connection schema.

SHT11 measures and sends these data in digital form to microcontroller since it has built-in analog-to-digital converter (ADC). SHT11 uses a bidirectional (two-wire) interface for data communication with microcontroller.

2.3 Driver and Fans

The frequency converter arranges the output voltage for the fans used in this study according to the analog signals coming from the control card. In developed EFLCS, YASKAWA V1000 model inverter/driver was used as frequency converter to drive the fans. This driver features 1.5 kW, single phase, 220-V 50 Hz input and 200-V output. Figure 4a shows the YASKAWA V1000 inverter/driver, and Fig. 4b shows the control inputs.

YASKAWA V1000 inverter/driver’s pulse width modulation (PWM) is V/f (voltage/frequency) -controlled, and the output frequency is between 0 and 400 Hz. The output voltage waveform has a constant amplitude. Variable output voltage is obtained by switching the source voltage at regular intervals. The output voltage is determined by adjusting the amplitude of the signal in each cycle. The fans were driven by this output voltage. Each fan used in developed EFLCS is 120 cm \(\times \) 120 cm \(\times \) 38 cm and operates with 0–220 V AC. Figure 5 shows one of the fans driven by the driver used in the design.

Fig. 5
figure 5

The fan

Full size image
Fig. 6
figure 6

PWM signals with different duty cycles and sample switching circuit [44]

Full size image
Fig. 7
figure 7

Control signal for PWM level 50

Full size image
Fig. 8
figure 8

a The driver’s output signal for PWM level 50 and b average of the driver’s output signal for PWM level 50

Full size image

The frequency converter receives the control signal from the control card of developed EFLCS in the range of 0–10 Vdc. The values of voltage, current and frequency that correspond to control signal value are produced as outputs and also can be monitored via this device. In this way, precautions can be taken against undesirable situations that may occur in the system.

2.3.1 Obtaining PWM Control Signal

PWM is a way of obtaining energy by consecutive pulses instead of a continuously changing (analog) signal. This pulse width is increased or decreased to regulate the energy flow. Figure 6 shows PWM signals with different duty cycles and sample switching circuit.

The control signal 0–10 Vdc required to control the driver was controlled by adjusting the pulse width of the voltage from 10 Vdc source. This adjustment or setting is implemented by triggering the IRF52N MOSFET through 0–5 Vdc output of microcontroller.

The PWM range of microcontroller is set to be between 0 and 255. In developed EFLCS, the input read by the sensors is evaluated by the program that makes use of FL approach and is loaded to microcontroller. Thus, the sharp values are obtained. These sharp values are assigned to the task period of the PWM signal to be generated by the PIC18F4620. This task period with a start value of zero is reassigned as soon as the sensors are read. Two examples are given below:

Fig. 9
figure 9

Control signal for PWM level 200

Full size image
Fig. 10
figure 10

a The driver’s output signal for PWM level 200 and b average of the driver’s output signal for PWM level 200

Full size image

For PWM level 50

The oscilloscope output of the control signal for PWM level 50 is shown in Fig. 7.

The oscilloscope image of the driver’s output signal corresponding to this signal (PWM level 50) is shown in Fig. 8a, and the oscilloscope image of the average of the driver’s output signal is shown in Fig. 8b.

For PWM level 200

The oscilloscope output of the control signal for PWM level 200 is shown in Fig. 9.

The oscilloscope image of the driver’s output signal corresponding to this signal (PWM level 200) is shown in Fig. 10a, and the oscilloscope image of the average of the driver’s output signal is shown in Fig. 10b.

2.4 Humidifier

The device that performs the cold humidification process is called UA. UA produces cold mist at approximately frequency range 1–2 MHz. The main material of UA is the asymmetric quartz crystals which, when the required frequency value is reached, cause mechanical vibrations to occur. After the main voltage is reduced by a transformer, it is converted to a 1.65 MHz frequency signal by an electronic circuit. This signal is transferred to a transducer mounted in a water reservoir. The transducer converts this signal to high frequency vibration. With the effect of the 1.65 MHz frequency signal, it changes its thickness by 1.65 million times a second. As a consequence of the high frequency vibration of the transducer surface and the momentum of the water due to the inertia of the water, this vibration of the water cannot keep up with it, which causes the water column to form in the transducer. During the transducer is shrinking in the negative direction, the water cannot follow this rapid movement and instantaneous vacuum occurs in the region above the transducer. Due to the vacuum, cavitation starts on the surface of the water and causes water particles to form under low temperature and pressure. During the positive expansion of the transducer, the water particles move rapidly toward the surface of the water column and collide with each other at high velocities. At the wave center of the sound waves on the water surface, the smallest water particles are broken down into 1-micron-sized mist particles. The produced mist is easily absorbed by the air and transported by the air current [45, 46]. Figure 11 shows the operation of the ultrasonic atomizer.

Fig. 11
figure 11

The operation of the ultrasonic atomizer [47]

Full size image

Blyss brand UA humidifier which has 220–240 V, 50 Hz, 32 W input values was used in developed EFLCS.

2.5 View of the Developed System

Figure 12 shows the developed EFLC system (unmounted) for CRDC. CRDC (dimensions 45 cm \(\times \) 1.5 m \(\times \) 1 m) used in this study and also the location of the equipment of the developed EFLC on this CRDC (mounted) are shown in Fig. 13.

Fig. 12
figure 12

Developed EFLC system (unmounted)

Full size image

As the sensor 5 is located near the evaporator, it cannot be seen in Fig. 13.

Fig. 13
figure 13

CRDC with developed EFLC system (mounted)

Full size image
Fig. 14
figure 14

Screenshot of the interface program

Full size image

2.6 Software of the System

The developed control system (EFLCS) together with PIC codes that is written to get information from sensor and to activate the actuators was transferred to microcontroller using PIC-C language with Hi-Tech PIC compiler. The interface software that displays the input–output values instantly on the computer screen was developed using Microsoft Visual Studio Program C# language.

2.6.1 Software of the Interface

The interface was designed to view and record the levels of both the steam of UA and the speeds of the fans that are controlled by developed EFCL. The interface is provided via the Universal Synchronous/Asynchronous Receiver/Transmitter (USART) RC6 port for communication with microcontroller. A screenshot of the designed interface program is shown in Fig. 14.

2.6.2 Software of Microcontroller

The inputs and outputs of the entire system are controlled by this software which is developed by using PIC-C language and transferred to microcontroller.

2.7 The Algorithm

The values (humidity and temperature) from the SHT11 sensors were used as inputs, and speed level of 3 simultaneously operated fans and steam level of UA were used as outputs.

As previously stated, UA is operated in levels (0–10 steps) depending on the average of the temperature and humidity values from the four sensors placed in the upper and lower parts of CRDC. Therefore, two inputs (average temperature and average humidity) and one output (steam level) EFLCS were designed for UA.

On the other hand, the speed levels of the fans depend on the temperature difference and average temperature values. The average temperature value is the average of the temperature values read from the four sensors placed in the upper and lower parts of CRDC. As mentioned previously for efficient air conditioning system, the difference between the average temperature of the cooled area and temperature of cooler liquid inside the evaporator must not exceed \(5\,^{\circ }\hbox {C}\). Due to this reason, the temperature difference between evaporator temperature and the average of the temperature values read from the four sensors placed in the upper and lower parts of CRDC is used to determine the current speed level values of the fans. Therefore, two inputs (average temperature and the temperature difference) and one output (fan speed level) EFLCS were designed for fans.

Both the code necessary for microcontroller to receive the information from the sensors and EFLCS code to operate the steam level of UA and speed levels of fan in accordance with this information were written and loaded to microcontroller. As a final addition, microcontroller was also loaded with the codes necessary for on–off control of the compressor between 3 and \(5\,^{\circ }\hbox {C}\) to keep the temperature inside CRDC constant at \(4\,^{\circ }\hbox {C}\) through a relay. When the design of EFLCS was completed, experiments were made on CRDC for different values of humidity and temperatures.

Normally, a fuzzy system is performed in three steps: fuzzification: mapping the actual values of the variables in the fuzzy set, inference: determining the rules that should be used according to the rule base, defuzzification: determining a numerical output according to the result obtained in the Inference step [48]. In this study, Mamdani implication method was used as inference mechanism, and the weight average method, which is most commonly used for the fuzzy applications and is limited to symmetric output membership functions [49], is used for the defuzzification process.

In the following sections, only the fuzzy membership functions of the inputs and outputs and also the rule tables are mentioned. The inference and defuzzification processes were carried out according to the previously mentioned methods and loaded to microcontroller.

2.8 Steam Level Control Design of UA

In FL control of UA, the average value of the temperature in CRDC for any moment is the first input value and the average value of the humidity in CRDC for any moment is the second input value. EFLCS decides at what steam level of UA will be driven according to these inputs at related moment. Depending on this decision or calculated output value, microcontroller generates a PWM signal to drive UA to reach the calculated steam level of UA at related moment.

2.8.1 Fuzzification of Inputs

The symmetric triangular membership function was used for moderate temperature (MT), tolerable temperature (TT) and high temperature (HT), whereas the trapezoidal membership function was used for low temperature (LT) and excessive temperature (ET). The average temperature input was defined by five membership functions as shown in Fig. 15.

Fig. 15
figure 15

Membership functions of average temperature input

Full size image

The symmetric triangular membership function was used for low humidity (LH), tolerable humidity (TH) and moderate humidity (MH), whereas the trapezoidal membership function was used for too low humidity (TLH) and high humidity (HH). The average humidity input was defined by five membership functions as shown in Fig. 16.

Fig. 16
figure 16

Membership functions of average humidity input

Full size image

2.8.2 Fuzzification of Output

The symmetric triangular membership function was used for low UA level (LUAL), moderate UA level (MUAL) and high UA level (HUAL), whereas the trapezoidal membership function was used for too low UA level (TLUAL) and excessive UA level (EUAL). UA steam level was defined by five membership functions as shown in Fig. 17.

Fig. 17
figure 17

Membership functions of steam level of UA

Full size image

2.8.3 Rule Table

The rule table for steam level of UA is shown in Table 3.

Table 3 Rule table for steam level of UA
Full size table

2.9 Speed Level Control Design of the Fan

In FL control of the fans, the difference between the evaporator temperature and the average temperature inside CRDC for any moment is first input and the average temperature inside CRDC for any moment is second input. EFLCS decides at what speed levels of fans will be driven according to these inputs at related moment. Depending on this decision or calculated output value, microcontroller generates a PWM signal for the YASKAWA V1000 driver to produce an output voltage value from 0 to 10 V to drive the fans to reach the calculated speed levels of fans at related moment. So, all the fans run at the same required speed level to provide required air conditioning in CRDC.

2.9.1 Fuzzification of Inputs

The symmetric triangular membership function was used for low difference between average temperature and the evaporator temperature (LATET), moderate difference between average temperature and the evaporator temperature (MATET) and high difference between average temperature and the evaporator (HATET), whereas the trapezoidal membership function was used for too low difference between average temperature and the evaporator temperature (TLATET) and excessive difference between average temperature and the evaporator (EATET). The range for the difference between the average temperature and the evaporator temperature was defined by five membership functions as shown in Fig. 18.

Fig. 18
figure 18

Membership functions of difference between average temperature and evaporator temperature

Full size image

The second input used for fan speed control design is the average humidity. Since the fuzzification of average humidity was explained in “design of steam level control design of UA” section (Sect. 2.8), the tables and figures will not be mentioned here again.

2.9.2 Fuzzification of Output

The symmetric triangular membership function was used for low fan speed (LFS), moderate fan speed (MFS) and high fan speed (HFS), whereas the trapezoidal membership function was used for too low fan speed (TLFS) and excessive fan speed (EFS). The speed level of fan was defined by five membership functions as shown in Fig. 19. Since all fans will run with the same speed level, the value of this output will be used to drive all fans.

Fig. 19
figure 19

Membership functions of level of fan speed

Full size image

2.9.3 Rule Table

The rule table for speed level of fan is shown in Table 4.

Table 4 Rule table for speed level of fan
Full size table

3 Results and Discussions

For comparison aim, the records (temperature and humidity) taken from sensors at certain intervals (13.7 s) were transferred to the computer. By using these records, graphics were obtained by Originlab program. First, the records of classical system were kept and then the records of developed EFLCS were kept using the same CRDC. Experiments were carried out at \(23\,^{\circ }\hbox {C}\) outdoor temperature and 55% RH outdoor humidity. The comparative graph of EFLCS and traditional system regarding the instantaneous temperature control is shown in Fig. 20. First of all, equal and long-term operation was carried out to stabilize both systems. Then, both systems were provided to continue running for 850 s more. According to the test results, the maximum temperature of the traditional system is \(5.56\,^{\circ }\hbox {C}\) and the minimum temperature is \(3.2\,^{\circ }\hbox {C}\). For EFLCS, the maximum temperature is \(4.44\,^{\circ }\hbox {C}\) and the minimum temperature is \(3.69\,^{\circ }\hbox {C}\). Considering that both systems are set at \(+\,4\,^{\circ }\hbox {C}\), it seems that EFLCS has less temperature fluctuation than traditional system.

Fig. 20
figure 20

Comparative graph of EFLCS and traditional system regarding the instantaneous temperature control

Full size image

The comparative graph of EFLCS and traditional system regarding the instantaneous humidity control is shown in Fig. 21. The humidity control results were obtained on the same conditions as the temperature test. According to the test results, the maximum humidity and minimum humidity of traditional system are 61.81% RH and 57.45% RH, respectively. Regarding EFLCS, the maximum humidity is 81.33% RH and the minimum humidity is 78.57% RH. Considering that both systems are set at 80% RH, it seems that EFLCS has obtained less humidity fluctuation around the set value whereas traditional system even could not reach the set value.

Fig. 21
figure 21

The comparative graph of EFLCS and traditional system regarding the instantaneous humidity control

Full size image

The comparative graph of EFLCS and traditional system regarding the reaching temperature set point (\(+\,4\,^{\circ }\hbox {C}\)) is shown in Fig. 22. The time of reaching the set point was measured instantly for both systems starting at \(+\, 7\,^{\circ }\hbox {C}\). Traditional system reached set value at 191.8 s and EFLCS at 109.6 s. EFLCS reached set value 82.2 s earlier than traditional system.

Fig. 22
figure 22

The comparative graph of EFLCS and traditional system regarding the reaching temperature set point

Full size image

The comparative graph of EFLCS and traditional system regarding the reaching humidity set point (80% RH) is shown in Fig. 23. The time of reaching the set point was measured instantly for both systems starting at 55% RH. Traditional system could not reach the set value (could only reach 60% RH in 342.5 s), whereas EFLCS reached set value in 191.8 s.

Fig. 23
figure 23

The comparative graph of EFLCS and traditional system regarding reaching the humidity set point

Full size image

4 Conclusions

In traditional CRDC produced in Turkey, generally only temperature on–off control is carried out. In the pastry sector, especially, the moisture level required for daily produced products cannot be ensured, which leads to a short shelf life. In this study, EFLCS for CRDC was developed to make the shelf life of the cakes longer than it was before. Developed EFLCS kept CRDC temperature and humidity values required for cakes at approximately \(+\, 4\,^{\circ }\hbox {C}\) and 80% RH, respectively. In addition, this study contributed to the literature by adding FL to the system developed by the authors [50]. This study is a rare study in which PWM- and FL-based control system is used in the field of climate control. Additionally, UA is used for the first time to supply required humidity in CRDC manufactured in Turkey. Thus, a different approach to this area is presented in this study.

As mentioned previously for efficient air conditioning system, the difference between the average temperature of the cooled area and temperature of cooler liquid inside the evaporator must not exceed \(5\,^{\circ }\hbox {C}\). With developed EFLCS, this difference is kept under \(5\,^{\circ }\hbox {C}\) and the snow on the evaporator surface was tried to be minimized to increase efficiency of evaporator.

Within the scope of the study, UA could be developed and the compressor could be driven at different speeds. Since each one is the topic of different fields, available UA and on–off control system were used during the compressor’s operating periods. The authors of this work on the one side continue to work to drive the compressor at different speeds, on the other side experimenting to demonstrate the superiority of EFLCS’s energy consumption over as per traditional system.

In this context, three test forms (MIL/SIL/PIL) that are successful and cost-effective solutions for automotive and aerospace embedded systems [51] can be used in FLC algorithms to be developed.