Design of a Smart IoT-Based Control System for Remotely Managing Cold Storage Facilities

2.1. Cold Storage Rooms Description

The experiment was conducted using two similar cold storage rooms CSRs established at the Date Palm Research Center of Excellence (DPRC), King Faisal University (KFU), Saudi Arabia. One of these CSRs was remotely controlled and monitored by the designed cloud-based IoT system (IoT-BC), and the second was left with its traditional control. Figure 1 shows the main components of the CSR, which consisted of a thermally insulated cold room, evaporator unit, condensing unit, control unit, humidification unit, and defrost unit. The internal measures of the cold room were 5.8 m in length, 2.9 m in width, and 2.9 m in height.

The CSR walls and ceiling were made from polyurethane sandwich panels (PUSP) with 0.12 thickness. The PUSP was composed of three layers. The internal and external layers were two corrugated aluminum plates with a thickness of 0.01 m. The core layer was high-density polyurethane foam with a thickness of 0.10 m. The evaporator (type: EVAP.CAE3264 R417A, Carrier s.c.s. Co., Aubagne, France) is equipped with two fans 230 V/single-phase/60 Hz. The R417 A refrigerant (Type: CAE3264, −40/38 °C min/max) was used in the cooling system of the CSR. The condensing unit is equipped with a refrigeration compressor 3 phase, 380–420 V/3-phase/60 Hz, and the nominal operating current was 6.8 A (model DLEE-201-EWK, Emerson Climate Technologies Ltd., Mikulov, Czech Republic). The control unit consisted of one 3-phase breaker, two single-phase breakers, three contactors, a 24 h timer, and a digital temperature controller 230 VAC (model: XR06CX, Dixell, Pieve d’Alpago, Italy). The control unit was used to control the operating fans and CSR compressor. The CSR included three compact fluorescent light bulbs with a total power of 75 W. The temperature inside the CSR was set at the target temperatures before conducting the experiments using the controller supplied in the control unit of each CSR.

The humidification unit was an ultrasonic humidifier with the same descriptions as the designed ultrasonic humidifier by Mohammed et al. [9]. The ultrasonic humidifier was used to control the relative humidity (RH) in the CSR. This humidifier consisted of ten ultrasonic transducers, a water reservoir, an air fan, and a control unit. The ultrasonic transducers were installed at the central bottom position of the reservoir. The frequency of each ultrasonic transducer was 2600 kHz with a resonance impedance of 2 Ω, and the operating temperature range was 0 to 50 °C. The humidifier reservoir was a rectangular stainless-steel tank of 0.25 × 0.25 × 0.40 m. The fan flow rate ranged from 0 to 2.5 m³/min with a maximum power consumption of 22 W, and the rated inputted voltage was 230 V/60 Hz. The RH was adjusted by the high-precision humidity control unit (MH13001 with a humidity sensor of INS121, Shenzhen Electronic Co., Ltd., Shenzhen, Guangdong, China).

The accumulated frost on the surface of the evaporator coils hinders the distribution and conduction of the cooling capacity of the refrigeration evaporator, which finally impacts the cooling efficiency or deforms the evaporator. This problem leads to increasing electrical energy consumption and reducing the service life of refrigeration system equipment. Therefore, the benefit of the defrost unit is to defrost the accumulated frost to improve the efficiency of the system cooling, ensuring the quality of stored products in the storage room, saving electrical energy, and extending the service life of the cold storage components. Therefore, the electric heating type (220 V, 200 W) was used for the defrost unit CSRs. The heaters of the defrost unit were clamped directly to the evaporator coil and operated by the control based on defrosting temperature sensor data to prevent the coils from damaging by freezing water.

2.2. The IoT-Based Control System Architecture

The detailed workflow and communication among the different entities of the designed IoT-based control system (IoT-BC) are shown in Figure 2. The figure shows the modified cold storage room (MCSR) that was monitored and controlled remotely. In addition, a set of sensors and microcontrollers for collecting data were used to control and monitor the selected parameters. The communication between the different entities is as follows:

• The sensors periodically collected the data from different portions of the cold storage room.

• The collected data were uploaded by the considered microcontrollers to our private channel on ThingSpeak through Wi-Fi for control, monitoring, and alert purposes.

• The data analytics was accomplished on our ThingSpeak channel using MATLAB Analysis App. According to the code written on MATLAB Analysis App, the desired action/control is sent in real-time to the designated microcontroller to be applied to the cold storage room.

• The authenticated users can access our private ThingSpeak channel and monitor real-time through graphical interfaces to all the measured parameters.

• Finally, two alerts (email and SMS) are sent to a designated administrator in case of an emergency based on real-time data analytics. The email alert was done using the MATLAB Analysis App. The SMS alert was accomplished using the ThingHTTP App and our account on the Twilio website.

The main goal of the designed IoT-BC is monitoring some critical parameters, i.e., the atmosphere storage, controlling its work, and sending alerts in cases of emergency. Therefore, the IoT-BC monitored the internal and external atmosphere temperature, internal relative humidity (RH), Internal Carbon Dioxide concentration (CO₂), internal ethylene gas concentration (C₂H₄), internal ambient light, compressor temperature, electrical current intensity, and electrical energy consumption.

The IoT-BC controlled the humidification unit based on the relative humidity levels, the ventilation unit based on the CO₂ and C₂H₄ levels, and the control unit based on the internal atmosphere temperature, compressor temperature, and current intensity.

The alerting email and SMS were achieved for the following parameters:

• Internal temperature measured by the internal DHT22 sensor;

• Internal relative humidity measured by the internal DHT22 sensor;

• Internal carbon dioxide CO₂ measured by the Senseair-S8 sensor;

• Internal ethylene gas C₂H₄ measured by the MQ-3 sensor;

• Internal ambient light measured by the BH1750 sensor;

• Compressor temperature measured by the DS1822-PAR sensor;

• The applied current intensity measured by three current meters.

2.3. Sensors Calibration

The Food and Drug Administration (FDA) requires an accurate display for all devices that have an impact on the quality of the stored product by their calibrating over a specified interval of time. The equipment calibration is also important to facilitate the storage of many critical materials that demand a controlled modified atmosphere and security. In order to comply with the international benchmarks, tested and calibrated CSR rooms (including their cooling equipment) according to the Code of Federal Regulations Title 21 were used in this study [38]. The sensors used in IoT-BC have been calibrated to assure their accuracy as follows:

• To calibrate the temperature and humidity sensor DHT22 and the digital temperature sensors, a controlled atmosphere (temperature and RH) incubator (model: PC900h, Helmer Scientific Inc., Noblesville, IN, USA) was used. The incubator atmosphere temperatures and RH were set at different values and the observed values were compared with the values of the sensor’s readings.

• To calibrate the CO₂ sensor, the reading of the Senseair-S8 sensor was compared with the reading recorded by the indoor air CO₂ m/datalogger (model: Extech EA80, FLIR Commercial Systems Inc., Nashua, NH, USA) at a temperature of 5 C. A carbon dioxide cylinder containing 99.5% pure CO₂ was used to add different concentrations of CO₂ in a closed room.

• To calibrate the ethylene MQ-3 sensor, a fruit ripening gas box containing 99.95% pure ethylene (type: corrugated box of 96 Pc, MILAN JYOTI INTERNATIONAL LTD, Mumbai, India) was used. The gas was sprayed in a closed cold room at a temperature of 5 °C and the reading of the sensor was compared with the reading of a portable ethylene gas detector (model: PG-100-C₂H₄, AMBETRONICS ENGINEERS PVT LTD., Mumbai, India).

• To calibrate the light intensity BH1750 sensor, the reading of the sensor was compared with the reading recorded by the light intensity meter/datalogger (model: Extech EA33, FLIR Commercial Systems Inc., Nashua, NH, USA). Calibration was carried out using the LED light source with variable intensity of illumination in the cooling room at a temperature of 5 °C.

• To calibrate the current intensity Load_CellX03 with HX711 amplifier, power, and electrical energy consumption, the reading of the current sensor, calculated power, and calculated energy consumption were compared with the reading recorded by three-phase power and harmonics clamp meter (UNI-T, UT243, Uni-Trend Technology Co., Ltd. Dongguan, Guangdong, China) under different current intensities consumed by a heater coil.

2.4. Control and Alert

IoT-BC’s control and the alert system have been introduced to control the compressor, the ventilation unit, and the humidification unit. As well as sending alerts to a designated administrator based on the measures of some sensors. Moreover, sending an alert based on the light level inside the atmosphere storage. In comparison to the Supervisory Control and Data Acquisition (SCADA) system which is very similar to the designed IoT-BC. The SCADA system could remotely control huge industrial processes in real-time based on high-level software applications. Additionally, a trained user is needed to use the SCAD system. Thus, the SCADA system is most suitable for the high cost and huge industrial facilities, as it is very expensive. The SCADA system also needs huge computational resources to work properly. On the other hand, the designed IoT-BC could achieve the same objectives as the SCADA system with low cost for the small to medium-sized industrial facilities. Moreover, the designed IoT-BC requires a simple IoT microcontroller and sensors along with a free cloud platform. Additionally, the normal user can use IoT-BC’s interface as it is very simple and easy to use.

The alert in IoT-BC is an email and SMS for a designated administrator through our private channel on the ThingSpeak cloud platform. The configuration for sending an email is achieved using MATLAB Analysis App on ThingSpeak. Where a MATLAB code is written in the MTLAB Analysis App for each controlled parameter. Additionally, the configuration of sending an SMS using ThingSpeak and the Twilio website is shown in Figure 6. The figure shows the detailed configuration for sending an SMS as an alert by using the Twilio website and ThingHTTP App on our private channel on the ThingSpeak cloud platform.

2.5. Energy Consumption

The voltage, current intensity, and power factor for TCSR were measured using a real RMS digital power and harmonics clamp meter (UNI-T UT243, Sinotronics Co., Ltd., Guizhou, China). The clamp meters were wired to a laptop for real-time data logging. The following equations were used to estimate the power and energy consumption:

where ${\mathrm{P}}_1$ is the actual power consumption for a single-phase, ${\mathrm{P}}_2$ is the actual power consumption for 3 phases, ${\mathrm{V}}_{\left(\mathrm{L}-\mathrm{N}\right)}$ is the voltage (line to neutral), ${\mathrm{V}}_{\left(\mathrm{L}-\mathrm{L}\right)}$ is the voltage (line to line), $\mathrm{I}$ is the current intensity, and ${\mathrm{P}}_{\mathrm{F}}$ is the power factor (${\mathrm{P}}_{\mathrm{F}}=\cos \mathsf{\phi })$, ${\mathrm{P}}_{\mathrm{F} }\simeq 0.94$ in the experimental site, $\mathrm{E}$ is the total energy consumption, and T₁ and T₂ are the actual operating time.

2.6. Tested Date Fruits

The yield of thirty date palm trees (Khalas cv.) were selected for this study from the experimental farms of the DPRC at the Agricultural Training and Research Station, KFU, SA (Latitude: 25.267690° N, Longitude: 49.708162° E). All selected date palm trees were almost the same age (16 years) and uniform in growth and were subjected to the same agricultural treatments. The fruits of these palm trees were harvested at a fully ripened stage (Tamr). Immediately after harvesting, the fruits with good characteristics and free from insect damage were transported to the postharvest laboratory at the DPRC. The fruits were sorted and cleaned using dry air and then divided into two similar groups (The total weight was 800 kg). The fruits were packaged in 40 plastic crates with an internal size of 50 × 40 × 25 cm. Each crate has been filled with approximately 20 kg of date fruit. The crates were stored immediately after packing in the two Cold Storage Rooms CSRs, i.e., MCSR and TCSR. The MCSR and TCSR have the same storage temperature (5 °C), same size (482), and manufacturing materials.

2.7. Fruit Quality Measurements

The most important quality parameters of the stored date fruits were measured before and after cold storage to study the impact of a controlled atmosphere on fruit quality. Therefore, the following physicochemical characteristics were measured immediately after harvesting at the Tamar stage (full ripened stage with brown color) and during cold storage use. The tested fruits were randomly selected before and during cold storage from different locations in the MCSR and TCSR.

The fruit weights were measured using a digital balance (MSA225S, Sartorius Lab Instruments GmbH Co, Göttingen, Lower Saxony, Germany). Fruit length (FL) and diameter (FD) were measured using a digital vernier caliper. For determining the percentage of fruit weight loss, the fruit samples were weighed before cold storage to obtain the initial weight and after 1, 2, and 3 months to obtain the weight after cold storage. The following equation was utilized to calculate the fruit weight loss:

where FWL is the fruit weight loss (%), W₁ is the initial weight of the sample (g), W₂ is the weight of the same date sample after the target cold storage time (g).

The projected area (PA) of fruits was measured based on the image processing method. A digital camera was used with a fixed height of 50 cm, and a light-emitting diode (LED) source with a power of 20 W was used to adjust the lighting. First, the captured images were transferred to a laptop via a USB port; then, the images were processed using the open-source image-processing software (ImageJ/Fiji 1.46) to determine the required parameters [39].

The fruit volume (FV) was estimated using the gas displacement method in an isothermal system with a pressure of 200 kPa and a temperature of 22 °C. This method used a cylinder with compressed air, an empty cylinder, and a treatment chamber made from stainless steel with 0.01 cm thickness, a length of 100 cm, and an outer diameter of 40 cm [40].

The fruit density (FD) was determined by dividing the mass of the sample by its volume, which was determined by the gas displacement methods. While the bulk density of dates was determined by filling a container of 5000 cm³ volume with a date sample, the date sample was weighted to measure its mass. The bulk density of the sample was calculated by dividing the measured mass by 5000 cm³.

Fruit sphericity of was defined as the ratio of fruit volume to the sphere with a diameter equal to the major length of the fruit. The fruit sphericity was determined using the following equation [41]:

where Φ is fruit sphericity, l, w, and t are the length, width, and thickness of the tested fruit (mm), respectively.

Fruit hardness was measured using a texture profile analyzer meter (TA.XTplus, Stable Micro Systems Ltd., Godalming, Surrey, UK) with a cylindrical puncture probe with a 7 mm diameter at room temperature (about 22 °C). The probe traveling speed was 0.5 mm/s, and the puncture distance was 5 mm. The maximum force was recorded during the punching process to indicate the hardness of the stored date fruits.

The moisture content of fruits was determined by drying a sample of 150 g at 70 °C under a vacuum for 48 h using the vacuum-drying oven (LVO-2041P, Daihan Labtech Co., Ltd., Namyangju-si, Gyeonggi-do, Korea) according to the standard methods of analysis of AOAC [42].

The pH, total soluble solids (TSS), moisture content (MC), and water activity (Aw) of the stored fruits were measured according to standard AOAC analysis methods [42]. The laboratory pH meter (HI-99121, Hanna Instruments, Leighton Buzzard, Bedfordshire, UK) was used to determine the pH data. The laboratory refractometer (RFM 860, Bellingham & Stanley Ltd., Kent, UK) was used to determine the TSS data. The portable water activity device (Aqualab Series 3, Decagon Devices, Inc., Pullman, DC, USA) Aw was used to determine the Aw data. A portable electronic moisture balance (Model MOC-120H, Shimadzu Corporation, Kyoto, Japan) was used to determine the MC data.

The color parameters of the fruits were determined based on CIELAB color space using a Hunter Lab color meter (Quest-45/0 LAV, Hunter Associates Laboratory Inc., Reston, VA, USA). The color parameters were measured for each trial using 10 date fruits selected randomly. Chroma, hue angle, and the color difference between the measuring color parameters of fruit before and after storage time were calculated using the following equations:

where L is the lightness, a is the red/green coordinate, b is the yellow/blue coordinate, C is Chroma, h° is Hue angle (degree), a is redness, b is yellowness, and ΔE is a color difference.

2.8. Statistical Analysis

The statistical metrics were utilized to validate the accuracy of the used sensors by comparing the measured values by the sensors with the reference value simultaneously using standard instruments. These metrics included the determination coefficient, root mean square error, and index of agreement. The determination coefficient (R-squared correlation (R²) expresses the strength of the correlation between the measured data and the observed data. The following formula expresses the R-value:

where R is Pearson correlation, N is the number of data points, X is the observed data, and Y is the measured data.

Root mean square error (RMSE) compares the difference between measured values by the sensor and the observed values measured by the standard instrument. The following formula defines the RMSE:

where RMSE is the root mean square error, X_i is the observed data, Y_i is the measured data, and N is the number of data points.

The index of agreement (IOA) describes the ratio of the mean square error and the potential error. The IOA varies between 0 and 1, which the agreement value of 0 indicating no agreement at all and 1 indicating a perfect match.

where IOA is the index of agreement, X_o is the observed data, Y_m is the measured data, ${\overline{\mathrm{X}}}_{\mathrm{o}}$ is the average data observed.

The statistical analyses of the data were conducted at a 0.05 significance level by the one-way analysis of variance ANOVA using the statistical analysis program of IBM SPSS version 24 (SPSS Inc., Chicago, IL, USA). In addition, the Tukey test was used to determine the least significant difference (LSD) between the experimental mean values at a 0.05 probability level.

3.1. Sensors Calibration

Some environmental factors, i.e., high-temperature levels, abnormal humidity conditions, sudden degradation, shocks, etc., can affect the accuracy of the measured values for the sensors used in MCSR. These factors let the sensors give inadequate measurements. Therefore, there is an urgent need to apply sensor calibrations that are subject to most of the mentioned factors to validate sensor measurements before use. In addition, the output signals of the sensors need to calibrate against the reference standard instruments for producing calibration curves that describe the responses of the sensors to the reference instrument data to ensure data acquisition quality [43,44].

Before deploying the sensors for the MCSR measurements, the sensors were calibrated against the suitable reference standard instruments. Table 1 shows the comparison between the measured values and the reference values for temperature, RH, carbon dioxide (CO₂), ethylene gas sensor (C₂H₄), light intensity, current intensity, power, and energy consumption using the standard statistical metrics of the Pearson correlation (R), determination coefficient (R²), root mean square error (RMSE), and index of agreement (IOA). This table shows that the calibration of the used sensors achieved the required accuracy within the target parameters based on the R, R², RMSE, and IOA, which had acceptable values for the used sensors.

The sensor calibrations have been applied for six sensors in IoT-BC, as shown in Figure 12 and Table 1. The observed values are plotted on the x-axis, and the measured values are plotted on the y-axis for each calibration curve shown in Figure 6. The details of each calibration curve are as follows:

• Figure 12a,b: These curves indicate the temperature and RH calibration. The measured temperature and RH using the DHT22 sensor have been validated with the observed measurements by the incubator. The R, R2, RMSE, and IOA are shown on the temperature and RH rows in Table 1, indicating a perfect match in the measured temperature and RH with the observed temperature and RH. The DHT22 sensors showed good performance at the various temperature and RH values and had good linear regressions that nearly overlapped the 1:1 line (y = x + 0). The linear regression was y = 1.0065x − 1.5962 for the temperature and was y = 0.8346x + 7.3535 for the RH.

• Figure 12c: The CO₂ measurements have been validated in this curve. The measured CO₂ using the Senseair-S8 sensor was validated with the indoor air CO₂ m/datalogger (Extech EA80). The values of R, R2, RMSE, and IOA are shown on the CO₂ row in Table 1, indicating a perfect match between the measured CO₂ and the observed CO₂. The Senseair-S8 sensors showed good performance at the various gas concentrations and had good linear regressions (y = 0.9389x + 0.004) that nearly overlapped the 1:1 line.

• Figure 12d: The C₂H₄ measurements have been validated in this curve. The measured C₂H₄ using the MQ-3 sensor has been validated with the observed measurements by a fruit ripening gas box containing 99.95% pure ethylene. The values of R, R², RMSE, and IOA are shown on the C₂H₄ row in Table 1, indicating a perfect match between the measured C₂H₄ and the observed C₂H₄. The MQ-3 sensors showed good performance at the various gas concentrations and had good linear regressions (y = 0.8445x + 0.0083) that nearly overlapped the 1:1 line.

• Figure 12e: This curve has validated the light measurements. The measured light using the BH1750 sensor has been validated with the observed measurements by the light intensity meter/datalogger (Extech EA33). The values of R, R2, RMSE, and IOA is shown on the light row in Table 1, indicating a perfect match between the measured light and the observed light. The BH1750 sensors showed good performance at the various light intensities and had good linear regressions (y = 0.9673x + 13.487) that nearly overlapped the 1:1 line.

• Figure 12f: This curve has validated the current intensity. The measured current intensity using the CST2 sensor has been validated with the observed measurements by the three-phase power and harmonics clamp meter (UNI-T). The R, R², RMSE, and IOA are shown on the current, power, and energy rows in Table 1, indicating a perfect match between the measured current intensity and the observed current intensity. The CST2 sensors showed good performance at the various current intensities and had good linear regressions (y = 0.9481x + 0.632) that nearly overlapped the 1:1 line.

3.2. Monitoring, Controlling, and Alerting Results

Figure 13 presents the real-time acquired data based on the sensors’ measurements for the most important parameters, i.e., outside temperature and compressor temperature (a), MCSR temperature (b), MCSR RH (c), CO₂ concentration in the MCSR (d), C₂H₄ concentration in the MCSR (e), and applied electrical current and hourly energy consumption (f) for 3 months from 1 January to 1 April 2022.

Figure 13a shows the variations in the maximum temperature (Ta Max), minimum temperature (Ta Min), and average temperature (Ta Avg) temperature for the outside atmosphere in the study area and the average temperature of the MCSR compressor (Tc Avg). The outside temperature ranged from 12.13 to 32.03 °C and the average outside temperature was 20.32 ± 5.09. The average compressor temperature was 62.34 ± 10.91 ranging from 32 to 80 °C. From the collected data in this figure, it is obvious that the outside temperature impacts the compressor temperature. There was no significant difference between the MCSR and TCSR compressor temperatures. The compressor of the MCSR has not automatically turned off, and the designed IoT-based system has not sent any email or SMS as an alert because the temperature has not reached to maximum compressor temperature Setpoint (100 °C for 5 min), as shown in Figure 8.

Figure 13b shows the maximum inside temperature (Tc Max) the minimum temperature (Tc), and the average temperature (Tc Avg) in the MCSR. Each value in this curve describes the average temperatures of three different locations in the MCSR. This temperature was controlled based on the flowchart shown in Figure 7, where Min C-Setpoint is the minimum control setpoint equal to 4.5 °C, and Min A-setpoint is the minimum alert setpoint equal to 3 °C. The Max C-Setpoint and Max A-Setpoint are the maximum control and alert Setpoint that equal 5.5 °C and 7 °C, respectively. Figure 13b assures that the developed MCSR successfully achieved the target internal temperature ranges as controlled and alerted based on the setpoints, as shown in Figure 7. Unlike the TCSR the designed system successfully sent an alert when the CCAR temperature reached 7.21 °C on 10 January at 8:00, 7.30 °C on 13 January at 14:15, 7.32 °C on 1 February at 12:10, 7.24 °C on 11 February at 4:16, 7.12 °C on 26 February at 22:10, 7.11 °C on 6 March at 20:15, and 7.11 °C on 30 March at 14:05 based on the Tc Max A-Setpoint (7 °C). The system also successfully sent an alert when the internal de-creased to 2.82 °C on 3 March at 10:15 and 2.87 °C on 11 March at 06:05 based on the Tc Min A-Setpoint (3 °C).

Figure 13c shows the average RH inside the MCSR for three months. The internal RH in MCSR is controlled based on the flowchart shown in Figure 10, where RH Avg is the average RH inside the cold storage room, Min C-Setpoint is the minimum control setpoint (75%), Max C-Setpoint is the maximum control setpoint (80%), Min A-Setpoint is the minimum alert setpoint that equals 65%, and Max A-Setpoint is the maximum alert setpoint that equals 90%. The control of the humidification system in the MCSR achieved the target RH ranges, as observed in Figure 10. There was a significant difference between the RH in the MCSR and TCSR. Where a significant oscillation in the RH in the TCSR was observed with the passage of storage time. Figure 13c assures that the designed IoT-BC has successfully achieved the target internal RH ranges as controlled and alerted based on the setpoints, as shown in Figure 10. The system successfully sent an alert when the average RH reached 91.53% on 6 January at 22:00 and 91% on 1 March at 20:05 based on the RH Max A-Setpoint (90%). The system also successfully sent an alert when the average RH decreased to 64.38% on 1 January at 06:00, 63.07% on 1 January at 18:07, 64.72% on 3 February at 02:10, 64.35% on 3 February at 12:00, 63.21% on 6 March at 20:10, and 62.21% on 30 March at 14:20, according to the RH Min A-Setpoint (65%).

Figure 13d shows the average CO₂ concentration (Avg CO₂) inside the cold storage room. The CO₂ concentration inside the cold storage was controlled and alerted based on the flowchart shown in Figure 11a, where Max A-Setpoint is the maximum control and alert Setpoint that equals 0.045%. Figure 13d assures that the designed IoT-BC has successfully achieved the target internal CO₂ ranges as controlled and alerted based on the setpoints, as shown in Figure 11a. Unlike the TCSR, the designed system successfully sent an alert when the average CO₂ concentration reached to 0.0477% on 10 January at 08:00, 0.0459% on 13 January at 14:00, 0.0462% on 1 February at 12:10, 0.0476% on 11 February at 4:15, 0.0476% on 26 February at 22:02, 0.0469% on 6 March at 22:05, and 0.0465% on 30 March at 14:15 based on the CO₂ Max A-Setpoint (0.045%).

Figure 13e shows the average C₂H₄ concentration (C₂H₄ Avg) inside the cold storage room. The C₂H₄ concentration inside the cold storage room was controlled and alerted based on the flowchart shown in Figure 11b, where Max A-Setpoint is the maximum control and alert Setpoint of 0.02%. Figure 13e assures that the designed IoT-BC has successfully achieved the target internal C₂H₄ ranges as controlled and alerted based on the setpoints, as shown in Figure 11b. The system successfully sent an alert when the average C₂H₄ concentration reached 0.0212% on 27 February at 04:10 and 0.0213% on 1 March at 20:15 based on the C₂H₄ Max A-Setpoint (0.02%).

Figure 13f shows the average required current (I Avg) and the hourly electrical energy consumption for the MCSR equipment. The electrical energy is consumed by the refrigeration of date fruits, evaporator, and condenser fans, defrost heaters, humidification system, lights, and control system. The average current for the cold storage room is monitored based on the flowchart shown in Figure 9, where Max A-Setpoint is the maximum control and alert Setpoint that equals 10 A. Figure 13f assures that the developed IoT-BC has successfully achieved the current target ranges as controlled based on the setpoints, as shown in Figure 9. There was no significant difference between the MCSR and TCSR applied electrical energy consumption although the MCSR was supplied with the designed control and humidification systems. The MCSR compressor has not automatically turned off and the designed IoT-based system has not sent any email or SMS as an alert because the applied electrical current has not reached to maximum current Setpoint (10 A for one minute) and the three current meters of the compressor recorded the normal ranges from 4 A to 7 A during its operating, as shown in Figure 9.

Inanition, the designed IoT-based system has not sent any email and SMS alerts due to the internal light has not remained greater than the maximum light Setpoint (L is greater than or equal to 5 Lux) for 30 min, as shown in Figure 11c.

3.3. Impact of IoT-BC on Stored Fruit Characteristics

Table 2 displays the impact of controlling and monitoring the MCSR by IoT-BC under controlled RH of 75–80%, 5 °C temperature, and CO₂ and C₂H₄ gases on the most important quality parameters, i.e., fruit weight, weight loss percentage, length, diameter (FD), projected area, volume, density, sphericity percentage, hardness, pH, total soluble solids, moisture content, water activity, color parameters, and color difference of stored date fruits (Khalas cv.) compared to the TCSR under a same controlled temperature of 5 °C during cold storage times.

The data in Table 2 show a highly significant (p < 0.05) effect of the storage time on the most quality parameters of stored date fruit in the TCSR compared with the MCSR. Comparing TCSR and MCSR after 3 months of storage time, date palm fruits stored in TCSR had significantly (p < 0.05) lower fruit length, volume, hardness, pH, and color parameters (b, h, and c). The fruit’s total soluble solids, moisture content, water activity, and color lightness (L) were significantly increased when date palm fruits were stored at different duration in the TCSR and the MCSR. There was no significant difference between the mean values of fruit weight, diameter, projected area, density, sphericity, and color parameter (a). There was no significant weight loss and color difference for the stored date fruit in the MCSR. On the contrary, there was a significant difference in weight loss and color difference for dates stored in the TCSR.

Food products must be properly stored in some cold storage facilities to prevent spoilage and save their nutritional value. At least, there should exist an effective control and monitoring for the temperature and humidity inside the cold storage facilities. This motivates the need for integrating IoT with the cold storage facility. The IoT-enabled cold storage monitoring can regularly and successfully record, monitor, and support the conditions inside the cold storage. It is also urgent to ensure that the cold storage room temperature never exceeds the optimal temperature based on the kind of product stored in the room.

The most critical parameters that affect the shelf life of the stored products in cold storage are temperature and RH [45,46]. The fruits of date palm cv. Khalas stored at 5 °C and 80% RH did not show insect activity or degradation in fruit quality and enhanced their shelf-life [47]. Soft and semi-soft date fruits can be preserved by storing them at 5 °C or lower, preventing the development of various pests [48]. However, the stored product is subject to partial weight loss during cold storage in conventional systems due to its surface drying in addition to the deterioration of the quality of the fruits during storage due to the lack of accurate control of the temperature and RH in the internal atmosphere of cold refrigerating rooms.

The morphological and biochemical characteristics of date palm fruit, such as size, color, hardness, acidity, sweetness, and moisture content, are used to assess its marketability [49]. The results found that fruit weight, size, and projected area were maintained better after three months of storage at MCSR, where the RH was remotely controlled. These parameters were significantly reduced at TCSR, where RH oscillated. Previous studies on date palm fruit storage indicated a reduction in fruit weight and size with time [4,50,51]. The management of storage temperature was emphasized in these investigations. However, the current study focused on storage temperature and RH. As a result, the fruits stored at MCSR did not reduce weight or size after three months of storage.

Similarly, the fruit weight loss was also minimal after three months of storage at MCSR compared to TCSR. The increase in respiration rate, induced by reduced RH in the air surrounding the fruit, is usually responsible for the decrease in fruit weight [52,53]. The fruit weight loss was higher with the increase in storage temperature in apples; however, higher RH reduced the weight loss [54]. The ultrasonic humidifier substantially impacted the weight loss of date fruits after six months of cold storage [9]. After three months of storage at MCSR, maintaining the RH retained fruit weight in the current study.

Similarly, the date palm fruit firmness was significantly reduced at TCSR compared to MCSR after six months of storage. Fruit tissue softening is linked to moisture loss and turgor pressure reduction [55]. In the present study, maintaining a high RH level equilibrated moisture content, which kept the fruit firm. Nectarine fruits have reduced weight loss and maintained higher fruit firmness during cold storage under 85–90% RH [56]. On the other hand, apple fruits kept at 95% RH lost their firmness very slowly during storage [54]. Similarly, high RH during storage-maintained firmness in zucchini fruit [57].

However, the fruit pH of date palm cv. Khalas stored at TCSR or MCSR was reduced after three months of storage; however, it was lowest at TCSR. The fermentation activity of microorganisms during storage is considered to produce organic acids, leading to decreased fruit pH [58,59]. Similar results were reported after fruit storage in cv. Khalas [16] and cv. Stamaran [60]. According to the present study, the TSS of date palm cv. Khalas stored for three months increased at both storage conditions, i.e., 15.96% at TCSR and 10.88% at MCSR. The leading cause of the increase in TSS would be the enzymatic conversion of large polysaccharides into small sugars [4]. According to Radi et al. [59], the increase in TSS of date fruits during storage may be due to microbial and enzymatic activity that converts high molecular weight molecules to low molecular weight compounds. Pomegranate fruit had a considerable rise in TSS content during storage, linked to moisture loss and increased sugar concentration in the fruit [61,62]. These results coincide with Aleid et al. [63], who stored date fruits (cv. Khalas) at 5 °C for 12 months and found that fruit pH declined while TSS and moisture content increased. Similarly, TSS increased in cv. Mazafati fruits are stored at 4 °C for 180 days [51]. Pomegranate fruits stored at 5 °C and 92% RH had significantly reduced weight loss and maintained TSS [64]. Our study also showed that the fruits were stored at TCSR and MCSR for three months, and the moisture content of the fruit increased by 30.85 and 18.85%, respectively, compared to the fruits before storage. Comparing the moisture content of fruits stored at TCSR and MCSR after three months, 16% less moisture content was determined in fruit stored at MCSR. Similarly, water activity was 14.44% less after three months of storage at MCSR compared to TCSR. Mohammed et al. [9] found that fruits of cv. Khalas stored at 5 °C and 80% RH maintained moisture content and water activity compared to traditional storage facilities. Another study revealed that after five months of storage, the moisture content in cvs. Majhoul and Boufeggous did not change considerably, whereas TSS increased in cv. Majhoul [65]. The decrease in moisture and water activity could be accounted for by the evaporation of fruit water caused by the relatively high temperature and moderate RH [66].

Apart from the values of CIE L, all other fruit color attributes (a, b, h°, and C) decreased during the storage time; however, these attributes (a, b, h°, and C) were significantly higher at MCSR when compared with TCSR color values. The decrease in h° was also reported in tomatoes during storage [67]. In cherries, the h° decreased under both cold and ambient temperature storage [68]. In the present study, the h° difference between unstored and stored fruits for three months at MCSR was non-significant, also reported in strawberries where storage temperature and RH did not affect h° [69]. A significant difference in color attributes was also observed in apple cv. Granny Smiths were stored for 4 months at 2 °C and 90% RH [45]. On the other hand, pomegranate fruits stored at 5 °C maintained their color better throughout the storage period [64].