Enhanced Sensor-Based Automatic Fire Suppression System for Residential Kitchen Safety ^†

Abstract

Fire outbreaks, whether caused naturally or unintentionally, pose serious threats to safety, especially in household environments such as kitchens. Common triggers include overheated personal devices, electrical malfunctions, and unattended cooking appliances. This study aims to develop and enhance an automated fire suppression system designed specifically for residential kitchen settings. The system integrates multiple sensors, photoelectric, ionization, and flame detectors, paired with an Arduino microcontroller to ensure accurate detection and timely activation of a servo mechanism that triggers either a Class A or Class K fire extinguisher. Through controlled testing using both solid and liquid combustible materials, we examined key variables, including sensor placement, height, and nozzle angle. The results from 15 trials per session revealed a correlation coefficient exceeding 0.90 between detection time and distance and the significance level of an analysis of variance of less than 0.05, indicating that increased distance significantly affects response time. The percent error remained below 6.7% across all tests, with strong correlations above 0.8 between combustible material type and the corresponding extinguisher class. This research contributes to the advancement of intelligent fire suppression systems by enhancing detection accuracy, reducing false triggers, and optimizing efficient sensor configurations for residential safety.

1. Introduction

Fire outbreaks in residential kitchens continue to pose a persistent safety hazard, primarily due to unattended cooking and the presence of both liquid and solid combustibles. The first phase of this study was to develop a servo-automated fire suppression mechanism capable of distinguishing between fire types using smoke and flame sensors, deploying either Class A or Class K extinguishers accordingly. The Centers for Disease Control and Prevention of the United States explains the importance of fire, which can reach temperatures of up to 160 °C, causing skin damage and health dangers for humans. Inhaling such particles from the smoke can cause heavy breathing, coughing, chest pain, and other symptoms [1]. The Philippines alone experiences major outbreaks of fire, which are primarily caused by electrical fires, the use of paper and combustible materials, and kitchen fires. As reported, some are left scarred because many houses lack automated fire suppression systems capable of suppressing the fire when it occurs. Almost 30% of people in America do not have their homes equipped with a suppression system, and 32% do not have an escape plan in place in case a fire breaks out in their homes [2].

Kitchen fire outbreaks are commonly caused by liquids or solid combustibles present in the area, such as kitchen oil or cooking oil used during cooking, and solid combustibles near the kitchen stove [3]. Several restaurants are equipped with automated fire suppression systems that can reduce the impact of a fire outbreak within the area. However, many ordinary people who use kitchen stoves for cooking have not installed an automated suppression system in their household, and such combustible materials may pose a risk to various individuals in protecting their homes.

The objective of this study is to analyze combustible materials, both solid and liquid, in kitchen fire scenarios using an automated fire suppression system designed to contain smoke and fire within the affected area. The study aims to assess whether the use of specific classifications of fire extinguishers—Class A for solid combustibles and Class K for liquid combustibles—can effectively reduce suppression time and minimize false triggers. Furthermore, various testing procedures be conducted to evaluate the accuracy and responsiveness of the system based on sensor placement, including height and distance, to determine the optimal configuration for maximum effectiveness in fire detection and suppression.

An automated fire suppression system is integrated with an Arduino microcontroller, which functions as the central processing unit for interpreting data received from smoke and fire sensors in the kitchen environment. The study emphasizes the importance of sensor placement by testing various heights and distances to evaluate the system’s ability to detect and respond to fire incidents accurately. The research focuses on the application of Class A and Class K fire extinguishers, specifically targeting solid and liquid combustible materials commonly found in kitchens. The experimental setup features a traditional Filipino kitchen design, complete with a standard stove equipped with a pan to simulate realistic fire scenarios [4].

2. Review of Related Literature

Fires originating from the kitchen are among the most common causes of residential fires in the Philippines. It often results from cooking oils, combustibles, or unattended cooking. Existing suppression systems are commercially available but often fail to provide sufficient coverage for various fire classes.

An early fire detection system prototype was developed to gather critical information, including location, temperature, gas content, humidity, and pressure. According to the study, the system can detect the gas content in 10 and 30 s for fires, with a maximum mounting height of 3 m. Aside from this, an ESP32 sensor was used to track and log the data, and it notify the user via email through its mobile application [5].

In 2023, an autonomous firefighting robot was designed to detect, classify, and extinguish two classes of fire: Class A and Class B. They used cloth, alcohol, wood, and gas to determine the response time indoors and outdoors. For the response time indoors, the values are 18.1 for cloth, 18.7 for alcohol, 18.2 for wood, and 19.3 for gas. For outdoors, the response times are 25.7, 26.2, 25.7, and 26.6, respectively [6].

3. Methodology

3.1. Conceptual Framework

One of the system’s inputs is the fire from the kitchen, and the other is the smoke it causes. There are five processes involved in the system: (1) The testing of different-degree angles and the best height of the nozzle to sufficiently extinguish the fire coming from the kitchen. (2) Determination and calibration of the best possible combination of fire and smoke sensors to be used for the system. (3) Passing of data from the smoke sensor through the Arduino (Arduino IDE ver. 2.3.2) and fire sensor, with the digital data from the Arduino being stored. (4) The Arduino captures the value coming from the storage and digitally encodes commands to be sent to the relay. (5) The relay opens or closes the servo motor, depending on the codes and commands of the Arduino microcontroller. The system outputs the chemical of the class K fire extinguisher into the target. The system’s outputs were interpreted accordingly (Figure 1).

Figure 2 illustrates the process by which the system triggers the fire extinguisher system. The initial input comes from the fire and smoke in a safe area, where it can be simulated with various types of fire. The infrared (IR) sensor, DHT sensor, flame sensor, and smoke detectors detect those initial values. The values are interpreted as temperature, fire, humidity, gas, and smoke. The Arduino captures those values and interprets them as parameters, allowing those values to be set. If the system can now detect that the set parameters are above, then it sends a command to the servo motor to activate the fire extinguisher and release the nozzle. The system also detects if the set parameters are being exceeded, then it sends a command to the servo motor to deactivate and stop the nozzle in the fire extinguisher.

3.2. Materials and Equipment

The Arduino serves as the central control unit for the automated fire suppression system. It is connected to two types of smoke detectors—photoelectric and ionization—alongside a flame sensor and an IR sensor, all of which are integrated into the board. Additionally, a motor is included in the system, functioning as the actuator responsible for deploying the fire suppressant (Figure 3).

Before conducting the fire tests, adherence to safety protocols was mandatory. In compliance with the Bureau of Fire Protection (BFP) of the Philippines, it is required to follow strict safety measures before and after any fire and smoke-related testing. The BFP provided guidance, and the tests were limited to a controlled number of combustible materials [7]. Personal protective equipment or PPE is required for all individuals to wear during the testing, comprising safety helmets, turnout gear, high-temperature resistance gloves, and boots (Figure 4).

3.3. Servo Motor and Stop

After obtaining the necessary values for the system, the system was encoded to trigger the servo motor and activate the suppression system. The parameters were set to capture and record the temperature, humidity, gas, fire, and smoke values, which were then written to a table for observation. Codes such as “valve open” the servo motor, while “valve close” is used to stop the suppression system (Figure 5).

3.4. Data Viewing

To view the data, users must connect to the WI-FI module of the Arduino; the service set identifier is “Fire Suppression,” and the password is “123456.” The user accesses the module of the system to read the values such as temp, hum, gas, fire, and smoke (Figure 6).

4. Result and Discussion

The data and parameters were gathered from testing various combustible materials, as well as the optimal height and distance required for the system to operate accurately. They highlight the importance of selecting the appropriate type of fire extinguisher for specific kitchen fire scenarios to ensure adequate suppression.

4.1. Calibration of Flame Sensor and Smoke Sensors

Table 1. Flame sensor calibration results.

Trial Number	Output Voltage (V)
1	2.15
2	2.01
3	1.95
4	2.34
5	2.12
6	2.15
7	2.01
8	2.17
9	1.97
10	2.23
11	2.34
12	1.95
13	2.01
14	2.11
15	2.21

,

Table 2. Photoelectric smoke detector calibration.

Trial Number	18 cm	30 cm	51 cm	75 cm	100 cm	144 cm
	Output Voltage (V)
1	2.70	2.45	2.11	1.47	0.95	0.51
2	2.61	2.15	1.99	0.88	0.99	0.57
3	2.72	2.21	1.88	1.18	0.57	0.60
4	2.70	2.50	2.05	1.03	0.61	0.49
5	2.57	2.55	2.10	0.79	0.61	0.50
6	2.65	2.25	1.96	1.39	0.75	0.55
7	2.80	2.40	1.77	1.12	0.77	0.51
8	2.72	2.50	1.60	0.96	0.70	0.35
9	2.72	2.21	1.57	0.75	0.65	0.40
10	2.50	2.35	1.32	1.17	0.61	0.45
11	2.62	2.45	1.88	1.25	0.52	0.51
12	2.57	2.52	1.89	1.30	0.43	0.62
13	2.70	2.51	1.77	1.15	0.43	0.65
14	2.51	2.52	2.11	0.83	0.51	0.43
15	2.62	2.20	1.95	0.96	0.55	0.39
AVG	2.65	2.38	1.86	1.08	0.64	0.50

and

Table 3. Ionization smoke detector calibration.

Trial Number	18 cm	30 cm	51 cm	75 cm	100 cm	144 cm
	Output Voltage (V)
1	1.75	1.35	1.01	1.12	0.52	0.25
2	1.08	1.35	0.98	1.14	0.53	0.23
3	1.42	1.30	0.95	1.04	0.48	0.19
4	1.18	1.25	0.92	0.93	0.41	0.25
5	1.60	1.20	0.89	0.80	0.27	0.16
6	1.10	1.05	0.95	0.87	0.35	0.21
7	1.15	0.98	0.99	1.08	0.35	0.14
8	1.33	1.28	1.02	1.11	0.50	0.17
9	1.27	1.12	1.01	0.76	0.44	0.12
10	1.55	1.07	0.88	0.95	0.39	0.20
11	1.33	1.18	0.85	0.89	0.59	0.18
12	1.50	0.92	0.82	1.00	0.44	0.22
13	1.75	1.02	0.85	0.82	0.46	0.11
14	1.70	1.02	1.00	1.13	0.25	0.15
15	1.55	1.15	1.01	0.90	0.44	0.24
AVG	1.42	1.15	0.94	0.97	0.43	0.19

present the calibration results of the flame sensor, photoelectric sensor, and ionization smoke detector based on the distance at which the system can detect its outputs. Each sensor underwent 15 trials, and the data shows that as the distance increases, the output voltage remains within the acceptable range recognized by the system, indicating consistent sensor responsiveness.

Fifteen trials are conducted to ensure that accuracy reading from the system is on point with the reading with extending distances from 18 cm, 30 cm, 51 cm, 75 cm, 100 cm, and 144 cm. The distances given are provided by the smoke detector manufacturer guidelines, of which the nearest distance that the sensor can read is about 18 cm and anything less than that will not be read by the sensor. Furthermore, anything above 100 cm will not be read by the sensor, so testing the absolute maximum distance that the sensor can read would result in no error being captured in the data. Other distances such as 30 cm, 51 cm, and 75 cm are also part of previous study, indicating the effectiveness of the flame and smoke sensors on those particular distances and how they provided optimum results that the study will also use. Every other test below will also make use of the distances in ensuring accurate reading from the system (Table 2).

Table 4. Flame sensor response time (unit in seconds).

Trial Number	18 cm	30 cm	51 cm	75 cm	100 cm	144 cm
	Response Time (Second)
1	1.38	1.56	1.53	1.86	2.34	4.51
2	1.33	1.57	1.87	2.37	2.68	4.65
3	1.42	1.42	1.72	2.85	3.14	4.72
4	1.36	1.49	1.62	1.94	3.01	4.55
5	1.45	1.61	1.55	1.94	2.87	4.39
6	1.39	1.57	1.55	2.63	3.67	4.30
7	1.34	1.52	1.80	2.05	2.48	4.07
8	1.41	1.46	1.61	2.10	3.32	4.44
9	1.45	1.65	1.67	1.88	2.55	4.51
10	1.45	1.59	1.99	1.85	3.23	4.30
11	1.37	1.44	1.56	2.75	3.72	4.52
12	1.32	1.53	1.75	2.44	2.91	4.33
13	1.43	1.41	1.68	2.58	3.45	4.15
14	1.40	1.50	1.86	1.92	2.38	4.12
15	1.35	1.63	1.89	2.18	3.56	4.46
AVG	1.39	1.53	1.71	2.22	3.02	4.40

,

Table 5. Photoelectric smoke detector response time (unit in seconds).

Trial Number	18 cm	30 cm	51 cm	75 cm	100 cm	144 cm
	Response Time (Second)
1	1.73	1.75	1.82	2.32	2.75	3.50
2	1.45	1.88	1.82	2.78	2.88	4.12
3	1.82	1.75	1.94	2.45	2.56	3.88
4	1.29	1.71	1.60	2.95	2.47	4.56
5	1.58	1.51	1.70	2.30	2.71	4.23
6	1.67	1.62	1.70	2.62	2.65	3.95
7	1.39	1.95	1.77	2.53	2.52	3.50
8	1.91	1.59	2.01	2.84	2.99	4.61
9	1.77	1.74	2.12	2.41	2.77	5.20
10	1.61	1.80	2.00	2.72	2.43	4.71
11	1.33	1.68	1.92	2.96	2.95	4.03
12	1.29	1.54	1.58	2.58	2.53	4.80
13	1.50	1.52	1.49	2.49	2.92	4.67
14	1.85	1.50	1.66	2.89	2.82	3.62
15	1.54	1.58	1.68	2.73	2.76	4.94
AVG	1.58	1.67	1.79	2.64	2.71	4.29

and

Table 6. Ionization smoke detector response time (unit in seconds).

Trial Number	18 cm	30 cm	51 cm	75 cm	100 cm	144 cm
	Response Time (Second)
1	2.38	2.45	2.90	3.10	3.05	4.56
2	2.63	2.75	2.91	3.01	3.56	4.32
3	2.63	2.81	2.90	3.13	4.25	4.85
4	2.41	2.80	2.76	3.12	4.13	3.76
5	2.63	2.46	3.08	3.30	4.01	3.79
6	2.22	2.72	3.01	2.85	3.86	3.75
7	2.51	2.57	2.74	2.92	4.l7	4.21
8	2.65	2.67	3.11	3.15	3.92	4.51
9	2.29	2.54	3.11	2.86	4.02	4.13
10	2.40	2.39	3.03	3.06	4.30	4.75
11	2.30	2.88	3.15	3.11	4.41	4.36
12	2.28	2.77	3.22	2.90	3.19	4.50
13	2.61	2.90	2.75	3.14	4.17	4.23
14	2.65	2.93	2.90	3.04	3.07	3.95
15	2.49	2.79	3.10	3.33	4.36	4.18
AVG	2.47	2.70	2.98	3.07	3.90	4.26

show that as the distance increases, the detection time recorded by the Arduino system also becomes longer, which is unfavorable for an automated suppression system that aims for a rapid response.

4.2. Sensor/s Combination Test

The use of the Response Time P.E.% or response time percentage error will determine the optimal combination of the sensors needed for the system, with the accepted value being the last type of fire utilized per sensor combination (in this case type III fire) and the experimental value being the total average time of all types of fire used per sensor combination testing [5]. In addition to the individual testing of the sensor and detectors, the combination of the flame sensor with either the photoelectric or ionization smoke detector was tested and when all sensors and smoke detectors are present. The accepted value is based on the final fire type tested, while the experimental value represents the overall average across all fire types (

Table 7. Sensor and combination test.

Sensor	Sensor Response Time (Second)			Total Average Time	Sensor Response Time (Percentage Error %)
	Type I Fire	Type II Fire	Type III Fire
Flame sensor	1.44	1.72	1.52	1.56	2.631578947
Photoelectric smoke detector	1.52	1.89	1.52	1.6433333	8.114035088
Ionization smoke detector	1.65	1.85	1.95	1.81666667	6.837606838
Flame sensor and photoelectric smoke detector	1.52	1.57	1.67	1.59	4.790419162
Flame sensor and ionization smoke detector	1.52	1.72	1.85	1.69666667	8.288288288
Flame sensor with photoelectric and ionization smoke detector	1.39	1.62	1.45	1.48666667	2.528735632

). Utilizing all three sensors resulted in the lowest percentage error of 2.53%, whereas setups using only one or two sensors had higher errors. Therefore, it is recommended to use the complete set of sensors at a distance of 18 cm from the fire and smoke source to achieve optimal performance of the suppression system.

$$ResponseTimeP.E.\%={\displaystyle \frac{|AcceptedValue-ExperimentalValue|}{AcceptedValue}}\times 100\%$$

(1)

4.3. Combustible Material Pearson Correlation Test

Using Pearson correlation, we assessed whether there is a positive or negative relationship between the type of combustible material and its corresponding detection time and distance. Data were manually recorded at heights ranging from 25 to 100 cm while testing both solid and liquid combustible materials.

Table 8. Paper combustible correlational value.

25 cm		50 cm		75 cm		100 cm
Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)
18	2.411	18	2.215	18	2.609	18	3.191
30	2.512	30	2.404	30	3.109	30	3.378
51	2.841	51	2.757	51	3.325	51	3.600
75	3.151	75	3.171	75	3.392	75	3.795
100	3.303	100	3.599	100	3.808	100	4.554
144	3.584	144	3.861	144	4.715	144	5.834
Correlation Value	0.981	Correlation Value	0.982	Correlation Value	0.974	Correlation Value	0.971

,

Table 9. Plastic combustible correlational value.

25 cm		50 cm		75 cm		100 cm
Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)
18	1.847	18	2.333	18	2.379	18	2.491
30	1.969	30	2.718	30	2.422	30	2.625
51	2.215	51	2.981	51	2.695	51	3.087
75	2.391	75	3.018	75	3.061	75	3.767
100	2.819	100	3.370	100	3.286	100	4.134
144	3.698	144	4.266	144	4.453	144	4.607
Correlation Value	0.984	Correlation Value	0.974	Correlation Value	0.976	Correlation Value	0.984

,

Table 10. Cloth combustible correlational value.

25 cm		50 cm		75 cm		100 cm
Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)
18	1.723	18	1.775	18	1.969	18	2.494
30	1.865	30	1.731	30	2.237	30	2.572
51	1.943	51	2.433	51	2.410	51	3.165
75	2.202	75	2.546	75	2.509	75	3.335
100	2.785	100	2.709	100	3.266	100	3.791
144	3.389	144	3.329	144	3.263	144	4.281
Correlation Value	0.981	Correlation Value	0.969	Correlation Value	0.942	Correlation Value	0.989

,

Table 11. Wood combustible correlational value.

25 cm		50 cm		75 cm		100 cm
Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)
18	2.368	18	2.544	18	2.816	18	3.006
30	2.552	30	2.544	30	2.976	30	2.989
51	2.893	51	2.991	51	3.224	51	3.054
75	3.091	75	3.015	75	3.771	75	3.173
100	3.296	100	3.351	100	3.765	100	4.093
144	3.617	144	4.206	144	5.594	144	5.823
Correlation Value	0.985	Correlation Value	0.975	Correlation Value	0.950	Correlation Value	0.917

,

Table 12. Coal combustible correlational value.

25 cm		50 cm		75 cm		100 cm
Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)
18	2.977	18	3.064	18	3.177	18	3.284
30	3.181	30	3.267	30	3.375	30	3.867
51	3.377	51	3.425	51	3.626	51	4.115
75	3.625	75	3.875	75	4.084	75	4.870
100	4.027	100	4.207	100	4.681	100	5.286
144	5.270	144	5.437	144	4.989	144	5.769
Correlation Value	0.972	Correlation Value	0.981	Correlation Value	0.985	Correlation Value	0.977

and

Table 13. Kitchen oil combustible correlational value.

25 cm		50 cm		75 cm		100 cm
Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)	Distance (cm)	Detection Time (Second)
18	1.433	18	1.559	18	1.772	18	2.251
30	1.647	30	1.695	30	1.805	30	2.356
51	1.988	51	2.344	51	2.045	51	2.614
75	2.063	75	2.851	75	2.529	75	2.746
100	2.810	100	2.991	100	2.965	100	3.550
144	4.293	144	3.687	144	4.307	144	4.599
Correlation Value	0.965	Correlation Value	0.982	Correlation Value	0.977	Correlation Value	0.972

display the results of testing various combustible materials at different system heights. At 25 cm—the closest distance—the detection time was the quickest, with the highest correlation value. This indicates that as the distance increases, the detection time also becomes longer. Among the materials tested, liquid combustibles had the fastest detection time, followed by solid combustibles.

4.4. Final Kitchen Fire Combustible Testing

A kitchen area was used for the test, along with the use of PPE and safety personnel. The area must be cleared of any debris so that no fire can break out (Figure 7).

Table 14. Nozzle response time unit in seconds (Class A).

Scenario of Kitchen Fire	1	2	3	4	5	6	7	8	9	10	Nozzle Response Time (Percentage Error, %)
Kitchen oil pan fire	1.87	2.11	1.98	2.42	1.73	2.30	1.79	2.59	2.03	1.92	8.020833
Combustibles near stove (paper)	1.45	1.80	1.85	2.00	1.95	1.60	1.84	1.62	1.92	1.88	4.734043
Combustibles near stove (plastic)	1.56	2.00	2.05	1.89	1.78	1.92	1.83	1.74	2.11	1.98	4.747475
Combustibles near stove (cloth)	1.73	2.14	2.01	2.06	1.67	2.21	1.92	2.00	1.81	2.08	5.625
Combustibles near stove (wood)	1.78	1.92	1.94	2.18	1.75	2.19	1.63	2.21	1.71	1.94	5.637255
Intended stove on	--	--	--	--	--	--	--	--	--	--	N/A

,

Table 15. After suppression temperature unit in degree Celsius (Class A).

Scenario of Kitchen Fire	1	2	3	4	5	6	7	8	9	10	After Suppression Temperature (Percentage Error, %)
Kitchen oil pan fire	28	29	30	32	34	35	32	31	33	35	6.66666667
Combustibles near stove (paper)	29	30	33	27	29	33	35	31	33	30	3.00
Combustibles near stove (plastic)	29	29	32	31	30	32	35	30	33	32	4.33333333
Combustibles near stove (cloth)	29	31	34	33	35	32	33	30	29	28	4.66666667
Combustibles near stove (wood)	29	33	29	31	29	35	33	29	31	36	5.00
Intended stove on	--	--	--	--	--	--	--	--	--	--	N/A

and

Table 16. Fire suppression time unit in seconds (Class A).

Scenario of Kitchen Fire	1	2	3	4	5	6	7	8	9	10
Kitchen oil pan fire	1.87	2.01	2.14	2.57	2.11	2.36	2.65	2.70	2.67	2.89
Combustibles near stove (paper)	1.78	1.97	1.88	1.45	1.55	1.92	2.01	1.95	1.82	1.98
Combustibles near stove (plastic)	1.60	1.85	1.92	1.78	1.75	1.99	1.97	1.67	1.85	1.81
Combustibles near stove (cloth)	1.92	1.92	2.11	2.51	2.92	2.11	2.01	2.45	1.87	1.75
Combustibles near stove (wood)	1.85	2.21	2.10	2.25	2.20	2.42	2.69	2.23	2.01	2.45
Intended stove on	--	--	--	--	--	--	--	--	--	--

reveal that the percentage error is higher for liquid combustible materials when using a Class A fire extinguisher. In contrast, solid combustible materials, ranging from paper to wood, yielded faster suppression results with the Class A extinguisher.

Table 17. Pearson correlation coefficient of fire extinguisher Class A.

Kitchen Oil Pan Fire		Combustibles Near Stove (Paper)		Combustibles Near Stove (Plastic)		Combustibles Near Stove (Cloth)		Combustibles Near Stove (Wood)
Suppression Time	Temperature After Suppression	Suppression Time	Temperature After Suppression	Suppression Time	Temperature After Suppression	Suppression Time	Temperature After Suppression	Suppression Time	Temperature After Suppression
1.87	28	1.78	29	1.60	29	2.55	31	1.85	29
2.01	29	1.97	30	1.85	29	2.25	31	2.21	33
2.14	30	1.88	33	1.92	32	1.22	32	2.10	29
2.57	32	1.45	27	1.78	31	2.43	29	2.25	31
2.11	34	1.55	29	1.75	30	1.92	31	2.20	29
2.36	35	1.92	33	1.99	32	2.25	34	2.42	35
2.65	32	2.01	35	1.97	35	2.05	30	2.69	33
2.70	32	1.95	34	1.67	30	2.01	32	2.23	29
2.67	33	1.82	33	1.85	33	2.45	31	2.01	31
2.89	35	1.98	30	1.81	32	1.35	33	2.45	36
Correlation Value	0.647	Correlation Value	0.708	Correlation Value	0.717	Correlation Value	0.681	Correlation Value	0.669

shows that there is a direct relationship between suppression time and temperature after suppression; as one increases, so does the other. Therefore, achieving a faster suppression time is essential for maintaining a lower post-suppression temperature.

Analysis of

Table 18. After suppression temperature unit in degree Celsius (Class K).

Scenario of Kitchen Fire	1	2	3	4	5	6	7	8	9	10	After Suppression Temperature (Percentage Error, %)
Kitchen oil pan fire	28	29	29	29	34	33	30	34	32	31	3.00
Combustibles near stove (paper)	28	29	30	31	30	30	33	34	33	35	4.33333333
Combustibles near stove (plastic)	30	30	31	31	32	30	31	35	34	30	4.66666667
Combustibles near stove (cloth)	28	28	29	33	30	33	35	34	32	33	5.00
Combustibles near stove (wood)	29	30	29	36	36	32	32	35	33	27	6.33333333
Intended stove on	--	--	--	--	--	--	--	--	--	--	N/A

,

Table 19. Fire suppression time unit in seconds (Class K).

Scenario of Kitchen Fire	1	2	3	4	5	6	7	8	9	10
Kitchen oil pan fire	1.30	1.33	1.45	1.49	1.52	1.67	1.69	2.01	1.79	1.67
Combustibles near stove (paper)	2.18	2.01	2.11	1.98	2.32	2.45	2.60	2.61	2.71	2.50
Combustibles near stove (plastic)	3.14	3.45	2.97	3.25	2.75	3.61	2.89	3.03	3.30	3.01
Combustibles near stove (cloth)	2.72	2.91	3.52	3.15	3.72	3.75	3.71	3.88	3.61	3.72
Combustibles near stove (wood)	2.91	3.02	2.87	3.50	3.68	3.99	3.71	4.15	3.99	3.60
Intended stove on	--	--	--	--	--	--	--	--	--	--

and

Table 20. Pearson correlation value of fire extinguisher Class K.

Kitchen Oil Pan Fire		Kitchen Oil Pan Fire		Kitchen Oil Pan Fire		Kitchen Oil Pan Fire		Kitchen Oil Pan Fire
Suppression Time	Temperature After Suppression	Suppression Time	Temperature After Suppression	Suppression Time	Temperature After Suppression	Suppression Time	Temperature After Suppression	Suppression Time	Temperature After Suppression
1.45	28	2.18	28	2.89	30	2.72	28	2.91	29
1.55	29	2.01	29	3.12	30	2.91	28	3.02	30
1.90	29	2.11	30	3.55	31	3.52	29	2.87	29
1.73	29	1.98	31	3.01	31	3.15	33	3.50	36
1.48	35	2.32	30	3.15	32	3.72	30	3.68	36
1.95	36	2.45	30	2.75	30	3.75	33	3.99	32
1.60	30	2.60	33	3.22	31	3.71	35	3.71	32
1.35	34	2.61	34	3.99	35	3.88	34	4.15	35
1.79	32	2.71	33	3.15	34	3.61	32	3.99	33
1.67	31	2.50	35	3.11	30	3.72	33	3.60	27
Correlation coefficient	0.745	Correlation coefficient	0.720	Correlation coefficient	0.700	Correlation coefficient	0.693	Correlation coefficient	0.557

shows that Class K fire extinguishers suppress both solid and liquid combustibles more quickly than the results shown in Table 14, Table 15, Table 16 and Table 17. Liquid combustibles are extinguished faster than solid combustibles with Class K extinguishers, where solid combustibles rank second. Furthermore, the Pearson correlation in Table 20 reveals that as suppression time increases, the temperature after suppression also rises.

5. Conclusions

The study results demonstrated that all integrated sensors, photoelectric, ionization, flame, and IR, provided acceptable output voltages across varying distances, with 18 cm identified as the optimal detection range. Among the sensor combinations tested, the integration of all four sensors yielded the lowest percentage error of 2.53%, proving to be the most reliable configuration. In testing various combustible materials, the results indicated that liquid combustibles were detected more quickly than solid combustibles. This is attributed to the denser smoke and more intense flames produced by liquids, which activated the sensors more rapidly. When analyzing the influence of system height, the optimal placement was found to be 25 cm, especially effective for detecting and suppressing fires from both solid and liquid combustibles. For instance, the use of a standard Class A extinguisher at this height suppressed liquid fires in an average time of 1.433 s and solid combustibles in under 2 s—ideal for automated fire response. Further testing compared Class A and Class K fire extinguishers. Class K was more efficient in suppressing both solid and liquid fires, resulting in faster nozzle activation, quicker temperature reduction to ambient room levels, and lower P.E. values overall.

Future studies are necessary to integrate an emergency notification feature, such as an SOS alert or automated text message system, to inform users when the suppression system is triggered. Enhancing the system with additional sensors that account for environmental or weather-related conditions can improve reliability. Expanding the system’s application beyond kitchen environments to more diverse settings could increase its utility. The inclusion of AI is also recommended, enabling the system to identify the specific type of fire and automatically deploy the appropriate fire extinguisher nozzle. Incorporating machine learning models could further enhance the system’s ability to distinguish between solid and liquid combustibles, thereby improving detection precision and suppression effectiveness.