SMART SOCKET WITH MULTIPLE INPUT AND LEARNING ENERGY MANAGEMENT SYSTEM.

Abstract

This invention relates to a smart socket multi-input learning energy management system that can be used in houses, apartments, detached houses, home offices, student dormitories, hotels, hostels, residences, offices, business centers, public buildings, educational institutions, universities, health facilities, hospitals, clinics, laboratories, industrial facilities, workshops, warehouses, production areas, commercial enterprises, shopping malls, smart building applications, smart city infrastructures, municipal service buildings, mass housing projects and all living and working areas where multiple electrical devices are used; its feature is; from a main body (100), from multiple socket inputs (101) located on the said main body (100), from CT current sensors (106) that detect the power consumption data of each socket input (101),The power signature is generated by digitizing the analog power consumption data obtained from CT current sensors (106) and from the power signature analog-to-digital converter and frequency analysis unit (107), the local AI processor, neural network module, relay control and Bluetooth module (108) which processes the generated power signature, identifies the connected device, learns user habits, performs inter-input dependency analysis and makes the decision to cut off the power, the independent relay (109) which independently switches the power supply of each socket input in line with the control signals generated by the local AI processor (108), the temperature and flame sensor array (110) which detects the temperature and flame information of each socket input, and the fire risk score bar indicator (103) which processes the temperature, current and anomaly data and creates the fire risk score.It consists of a WiFi / Bluetooth control panel (102) that displays system status information and fire risk information, and a mobile application notification output (111) that transmits status and warning information to the user device.

Description

1 TARIFF SMART SOCKET WITH MULTIPLE INPUT AND LEARNING ENERGY MANAGEMENT SYSTEM. Technological Field: 5 This invention is suitable for houses, apartments, detached houses, home offices, student dormitories, hotels, guesthouses, residences, offices, business centers, public buildings, educational institutions, Universities, healthcare facilities, hospitals, clinics, laboratories, industrial facilities, workshops, warehouses, production areas, commercial businesses, shopping malls, smart building 10 applications, smart city infrastructures, municipal service buildings, mass housing projects and in all living and working areas where multiple electrical devices are used The usable smart plug is related to a multi-input learning energy management system. State of the Art: 15 Efficient use of electrical energy occurs during standby mode. in order to reduce unnecessary energy consumption and increase electrical safety Various smart plug and energy management systems have been developed. Among the known techniques are: Fixed-time smart plugs, remotely controlled WiFi-based plugs, standby power breaker 20 Adapters, AI-powered smart plug solutions, and centralized smart home automation. These systems generally provide power to the devices using specific methods. opening and closing according to the rules or remote control by the user It aims to provide energy management and security functions in different ways. It is carried out using these methods. 25 Fixed-time smart plugs allow the user to set a predetermined time. Energy supply or cut-off operations are implemented according to their programs. These systems... It does not analyze user habits, therefore it cannot adapt to changes in usage patterns. This is not possible and requires manual updates of the programs. 30 Devices may malfunction if the user's daily routine changes. the power being cut off at certain times or remaining energized for an unnecessary period of time 2 This can be a topic of discussion. Similarly, remotely controlled WiFi power outlets can affect energy management. This is done via mobile applications and for each power cut operation. It requires manual intervention from the user. Also, these systems use the internet. Because it operates dependently on the connection, it automatically shuts down in case the connection is interrupted. Energy management functions may be disabled. 5 Standby power adapters typically operate on a single power outlet and It makes decisions to cut off power based on predetermined fixed power thresholds. This The solutions are unable to determine the type of connected device or its usage characteristics; different It is unable to adapt itself to the devices and only offers 10 entry-level devices. It is conducting an evaluation. Inputs in environments where multiple devices are used together. considering the usage relationships or interconnected work scenarios between them Since it is not taken, it is possible to completely eliminate unnecessary energy consumption. This is not possible. Also, the power consumption characteristics of different devices vary. In this case, systems may need to be restructured and incorrect power cuts may occur. decisions can be made. Even the well-known AI-powered smart plug solutions have various limitations. For example, the system described in Korean patent application KR20180029362A. It performs energy control by utilizing user location information and 20 It operates dependent on an internet connection and uses a single outlet instead of a multi-outlet base architecture. It is geared towards its structure. The EnAPlug system can learn energy consumption profiles. The socket base is not in the form factor; it analyzes the dependency relationships between inputs. It is unable to do so and does not constitute a commercially widespread solution. General smart home These systems require a central server infrastructure and a constant internet connection. 25 This increases installation and maintenance costs, creates a need for complex infrastructure, and each It does not allow the power outlet to operate independently. Furthermore, existing systems... None of them have device fingerprint recognition, user habit learning, or cross-entry functionality. dependency analysis, fire risk score calculation, and anomaly detection functions 30 integrated into a single multi-socket base, independent of internet connection. There is no structure in place where this is implemented. Therefore, the efficiency of standby energy... Managing it in this way, preventing wrong energy cut-off decisions, usage 3 adapting to habits, analyzing multi-device relationships, and fire Known techniques are important for identifying risks at an early stage. It has shortcomings. Description of the invention: 5 The invention utilizes local artificial intelligence algorithms operating on a multi-input power strip base structure. By continuously analyzing user habits, it saves energy in standby mode. It automatically manages consumption. Usage patterns may change over time. The ingenious system dynamically updates the learning process and eliminates any manual 10 Updating power cut-off decisions without requiring programming or reconfiguration. It is created according to usage habits. Thus, it eliminates the need for user intervention. Energy efficiency is continuously maintained without being noticed. The device used within the scope of the invention is a fingerprint identification mechanism, with 15 fingerprints associated with each entry. We determine the power consumption signature of the device using an FFT-based analysis method and different It enables automatic identification of devices. A new device is placed in the socket. When connected, the system can relearn the device, identifying which device is running each time it connects. It can determine with high accuracy and make power cut-off decisions based on the device's actual performance. By designing it according to the usage characteristics, the probability of incorrect cutting is significantly reduced. It reduces. One of the key advantages of the invention is the analysis of dependencies between inputs. It is the ability to accomplish this. The system enables collaboration between devices used together. By learning about their relationships, they can generate chain-reaction energy management decisions. For example, a 25 If the main device ceases to be used, then the other devices associated with it will also cease to be used. by enabling the power to be automatically cut off, something the user might forget about. It performs shutdown operations autonomously. This system saves energy. While increasing efficiency, it also helps prevent unnecessary electricity consumption. The invention also enables independent temperature and current monitoring for each inlet, resulting in... Using the collected data, we calculate the fire risk score and detect abnormal operation. 4 It can identify the conditions at an early stage. The risk level is within the defined limits. If the values ​​are exceeded, only the power to the relevant input is automatically cut off, and The user is informed via the mobile application. As a result, excessive electrical safety risks that may arise from heating and current anomalies fires that may originate from devices operating in standby mode are being reduced, especially 5 The probability is significantly reduced. All learning, analysis, and decision-making mechanisms of the invention require an internet connection. with local artificial intelligence algorithms running on the device without hearing This is done in such cases that the internet or WiFi connection is interrupted. Energy management and safety functions continue uninterrupted, user Since the data is not transferred to any external server, data privacy is fully ensured. It is protected. This structure, which does not require cloud infrastructure or subscription, is for business purposes. It also reduces costs. Developed in the form of a standard socket base, the system requires additional infrastructure, a central control unit, or It can be used without requiring complex installation. Independent relay for each input. Thanks to its control system, selective energy management can be implemented, and the system can be used from a single room to multi-story buildings. easily without the need for a centralized management infrastructure down to the buildings It is scalable. Thanks to these features, it can be used for residences, student dormitories, hotels, offices, 20 in different usage scenarios such as healthcare facilities, industrial areas and public buildings It provides practicality. The invention relates to energy management, device identification, input dependency analysis, and fire protection. Thanks to its design that integrates safety functions into a single socket base, the electrical system... 25 This contributes to reducing consumption, resulting in approximately monthly energy costs. It can provide savings of 10-20% and thanks to its one-time installation structure. It can recoup its investment costs in approximately 6–12 months. Local artificial intelligence. learning, device fingerprint identification, dependency analysis and integrated fire risk Bringing together scoring functions in the same system makes the invention a technical innovation. This constitutes one of the important innovations it has brought. Explaining the Figures: The invention will be described by referring to the attached figures, so that the features of the invention can be explained. It will be understood and appreciated more clearly, but the purpose of this invention is this obvious It is not about limiting it with regulations. On the contrary, the invention is defined by the accompanying claims in 5 all alternatives, modifications, and options that could be included within the defined area The aim is to cover their equivalences. The details shown are only for the present invention. It is shown to illustrate the preferred arrangements and both the methods shaping, as well as the rules and conceptual features of the invention, in the most useful way. It should be understood that they are presented to provide a readily understandable definition. These 10 in the drawings; Figure 1 shows a schematic view of the main body. Figure 2 shows a schematic view of the sensor. Figure 3 shows a schematic view of the system. 15 Illustrations that will help understand this invention are shown in the attached image. They are numbered and their names are given below. References Explanation: 20 100. Main body 101. Power outlets 102. WiFi / Bluetooth control panel and fire risk score indicator. 103. Fire risk score bar indicator 25 104. USB-A and USB-C output ports 105. Main power cable connection 106. CT current sensor 107. Power signature analog-to-digital converter (ADC) and frequency analyzer unit. 108. Local AI processor, neural network module, relay control and Bluetooth module 30 109. Independent relays 110. Temperature and flame sensor array. 6 111. Mobile app notification output Description of the Invention: The invention consists of a main body (100) and 5 located on the said main body (100). power consumption data for each socket (101) from multiple socket inputs (101) from the detecting CT current sensors (106), obtained from the CT current sensors (106) The power signature is created by digitizing analog power consumption data. power signature generated from digital converter and frequency analysis unit (107) By processing, it identifies the connected device, learns user habits, and analyzes inputs between 10... The local AI processor that performs the dependency analysis and makes the decision to cut off the power, from neural network module, relay control and Bluetooth module (108), local AI processor (108) Each socket input is powered according to the control signals generated by the system. from the independent relay (109) which switches the supply independently to each socket input temperature and flame sensor array (110) which detects temperature and flame information, temperature, 15 The fire risk score bar creates a fire risk score by processing current and anomaly data. from its indicator (103), displaying system status information and fire risk information Status and alert from WiFi / Bluetooth control panel (102) to user device. It consists of mobile application notification output (111) which transmits information. The invention is a power signature that separates power consumption data into its frequency components using an FFT algorithm. It has an analog-to-digital converter and a frequency analysis unit (107). The invention provides a local system that identifies a connected electronic device by storing power signature data in memory. 25 with AI processor, neural network module, relay control and Bluetooth module (108) is happening. The invention processes user's time-based usage data to analyze usage habits. the local AI processor, neural network module, relay control and Bluetooth module that determine (108) has. 30 7 The invention analyzes the time relationship between usage data from different power outlets. a local AI processor, neural network module, and relay that determine the dependency between inputs. It has a control and Bluetooth module (108). The invention evaluates temperature data and current consumption data together to identify anomaly data 5. It has a temperature and flame sensor array (110) that is formed. The invention is based on the control signals generated by the local AI processor (108) for each It has an independent relay (109) that switches the socket input separately. The invention transmits system status information and fire risk information wirelessly. This includes the mobile application notification output (111) which is transferred to the user device. The invention uses CT current sensors (106) to obtain power consumption data for each socket input. through which the perceived power consumption data is detected, the power signature is analog-digital 15 digitization by the converter and frequency analysis unit (107) and frequency decomposition of the generated power signature data into its components by the local AI processor (108) processed by identifying the connected device and the user's usage data. By processing the data, usage habits are determined, and usage patterns for different power outlets are analyzed. Determining the dependency relationship between the data, temperature and current data 20 anomaly information is generated by evaluating the results of the analysis. Control of independent relays (109) and the system status and risk created information is sent to the user's device via the mobile application notification output (111) It includes transmission. The invention converts power consumption data into frequency components using an FFT algorithm. It includes the separation step. The invention involves identifying a connected electronic device using a generated power signature. It includes step 30. 8 The invention explores the dependency between different power outlets using usage data from those outlets. It includes the step of determining the relationship. The invention provides anomaly information by evaluating temperature and current data together. 5 creation and control of independent relays (109) based on this information It includes the step. Detailed Description of the Invention: The invention allows for the use of multiple electrical or electronic devices within the same energy distribution system. It allows power to be supplied through the system, the power consumption of each device to be monitored individually, and the connected devices... the identification of devices according to their power consumption characteristics, user learning their habits locally, usage between different power outlets Identifying dependencies, preventing unnecessary energy consumption, and temperature, Detecting electrical risks based on flame, current and anomaly data. 15 Smart plug providing multiple inputs as a learning energy management system. It is being structured. All physical and electronic components of the invention are designed to be used in an external environment. 20 that protects against its effects and provides user access in an integrated manner. on the main body (100) or inside the main body (100) It is located. The main body (100) contains the electrical power distribution components, data processing units, sensing components, communication units, display components and is configured to carry the switching elements. On the main body (100), different electrical or electronic devices can be connected to the system. There are multiple socket inputs (101) that allow connection. Socket inputs (101), enabling the power supplies of connected devices to be monitored independently of each other and They are linked through separate energy channels in a way that allows them to be controlled. The current drawn by the device connected to each socket (101), operating time, and standby time are 30. The status and power consumption pattern are monitored separately. 9 The system is connected to the electrical grid via the main power cable connection (105). The main power cable connection (105) is made from the external power grid. the received electrical energy to the power distribution line located within the main body (100) It transmits. The electrical energy received from the main power cable connection (105) is independent. The relays (109) distribute the power to the socket inputs (101) and the electronic system 5 power supply for control, sensing, communication, display and data processing components It provides. USB-A and USB-C output ports (104) on the main body (100) It is located. USB-A and USB-C output ports (104), phone, tablet, portable 10 computers, accessories and similar electronic devices must be connected to a suitable electrical outlet. It provides power or charging. It has USB-A and USB-C output ports. (104) transmitted electrical energy, energy taken from the main power cable connection (105) by converting the voltage and current values ​​to those required by the relevant devices is provided. 15 The operation of the system is based on the connection of the main power cable (105) to the electrical grid. It begins with the connection. After electrical energy is transmitted to the system, the main The control and sensing units inside the housing (100) are energized, and the socket inputs are connected. (101) powered via independent relays (109), USB-A and USB-C output ports 20 (104) is made ready for use and the local data processing process is started. CT current sensor (106) associated with the power line of each socket input (101), It detects the current drawn by the device connected to the relevant power outlet. CT current The sensors (106) measure the power consumption at the sockets independently of each other. and generates analog signals regarding the instantaneous power consumption behavior of connected devices. The measurement process involves the device's active, switching, standby, and shutdown states. at specific time intervals or continuously, in a way that allows for its determination is being carried out. Analog power consumption signals obtained from CT current sensors (106), power signature The power signature is transmitted to the analog-to-digital converter and frequency analysis unit (107). Analog-to-digital converter and frequency analysis unit (107), converts analog signals to digital power It converts consumption data. The digitized data is broken down by time domain and frequency. Data processed in terms of area to represent the operating characteristics of the connected device. They are converted into clusters. Power signature analog-to-digital converter and frequency analysis unit (107), digitized It decomposes power consumption data into frequency components using an FFT algorithm. The initial operating current of the device is determined stably from the data broken down into its frequency components. current consumption, transient consumption variations, standby consumption, and load variations. Distinguishing electrical properties, such as behavior, are identified. These properties, when combined, form 10 By introducing these signals, a power signature is created that represents each connected device. The time and frequency dependence of the generated power signature is shown in the device power signature graph. (111) is represented on the device power signature graph (111), to a socket input. (101) Visual or numerical data of the current and power consumption characteristics of the connected device 15 It enables the device to be shown in its structural form. The device power signature graph (111) characteristic patterns, differentiating between devices and previously learned information It is used for comparison purposes with other devices. Distinguishing features obtained from the device power signature graph (111) indicate the local AI processor, 20 The neural network module transmits the relay control and Bluetooth module (108). Local AI processor, neural network module, relay control and Bluetooth module (108), belonging to connected devices It processes power signatures locally and performs device identification, learning user usage patterns, analyzing the relationships between data from different inputs It evaluates relationships and makes control decisions regarding energy management. 25 Local AI processor, neural network module, relay control and Bluetooth module (108), first power consumption data of the device connected to each socket (101) during the usage process It collects data over a specific period of time. Based on the collected data, the device can be turned on, Operating, waiting, and shutdown behaviors are learned, and the power representing the device is understood. The signature is saved in memory. Thus, the television, game console, computer, charger 30 devices such as desk lamps, sound systems, media players, and similar devices consume power. They are distinguished from each other according to their patterns. 11 Connecting a device different from the previously registered device to the same socket (101) In this case, new data obtained by the CT current sensor (106) shows the power signature. analog-digital converter and frequency analysis unit (107) re-analog is being processed. The power signature of the new device is compared with the registered power signatures and 5 If no match is found, a new device profile is created. The newly created profile... device profile, local AI processor, neural network module, relay control, and Bluetooth module (108) is saved to the system memory. Local AI processor, neural network module, relay control and Bluetooth module (108), 10 It also processes time-based usage data for each user. This data per device per day. at what times it is operated, how long it remains active, and when it is in standby mode. It is determined when it switches to this mode and after which usage cycles it is shut down. By evaluating daily, weekly and period-specific usage changes. User-specific usage habits are being created. 15 Local learning model if user habits change over time. It is being updated. The system treats previous usage behavior as a fixed rule. Instead of implementing it, it continues to process current usage data. Thus... usage hours of devices, working periods and usage between devices 20 Changes in these relationships are reflected in energy management decisions. Usage data of devices connected to multiple sockets (101), between inputs The inputs are compared within the scope of the dependency analysis scheme (112). Dependency analysis scheme (112), devices located in different socket inputs have the same or 25 representing links relating to their use in successive time intervals This diagram shows how the operation of one device affects the operation of another device. whether it is connected or another device after a device is turned off It is determined whether or not it continues to be used. For example, the television and the game console are connected to different power outlets (101). In this case, the relationship between the usage times of the devices in question is between inputs. 12 It is processed on the dependency analysis scheme (112). After the television is switched off repeated detection that the game console has not been used for a certain period of time If this is the case, the game console will have a usage relationship connected to the television. It has been determined that this is the case. After this relationship was discovered, the television was turned off. and when the specified time has elapsed, the power outlet of the game console will turn off. A decision is being made to discontinue feeding. Similarly, computers and peripherals, televisions and sound systems, media players. between the display device or other electronic devices used with it Time relationships were also analyzed using the input dependency analysis scheme (112) 10 This is determined by only one factor when the system makes a decision to cut off the power. Not only the power consumption of the devices but also the usage relationships between the devices are taken into consideration. is being received. Local AI and anomaly detection engine (113), CT current sensors (106), power signature 15 analog-to-digital converter and frequency analysis unit (107), local AI processor, neural network module, relay control and Bluetooth module (108) and temperature and flame sensor It evaluates the data obtained from the series (110) together. Local AI and anomaly Detection engine (113), learned consumption and temperature under normal operating conditions It identifies differences between patterns and real-time data. 20 Local AI and anomaly detection engine (113), unusual current surge, prolonged persistence High power consumption, unexpected changes in power signature, increased temperature at the power outlet, Situations such as flame detection or the simultaneous occurrence of these data are considered anomalies. It evaluates anomalies not only based on a fixed threshold value, but also on 25 according to the previously learned normal operating characteristics of the device in question is being carried out. Temperature and flame data for each socket input (101), temperature and flame sensor array (110) It is detected by. Temperature and flame sensor array (110), in socket inputs or 30 temperature occurring in electrical connection areas associated with power outlets 13 It monitors changes. The flame detection component detects electrical faults, overheating, or It enables the detection of flame formation associated with the onset of combustion. Data from the temperature and flame sensor array (110) are obtained from the CT current sensors (106) With the current data obtained, the local AI and anomaly detection engine (113) 5 It is being transmitted. The temperature increase occurs together with high current consumption, If the device deviates from its normal consumption pattern or a flame detection occurs, the relevant The risk level associated with the power outlet is being increased. Based on the evaluation of temperature, current, and anomaly data, the fire risk score is 10. The generated fire risk score is the fire risk score bar indicator (103) The fire risk score bar indicator (103) is shown to the user via the system. to enable the gradual monitoring of the risk level determined by It is being structured. If the fire risk score exceeds the predetermined safety level, the local AI and anomaly detection engine (113), local AI processor, neural network module, relay control and It creates a decision to cut off power via the Bluetooth module (108). the control signal to the independent relay (109) belonging to the socket input (101) where the risk is detected This is transmitted. Thus, while the power supply to the input at risk is cut off, the other socket 20 Work on the inputs is ongoing. Independent relays (109) supply power to each socket input (101) separately. It is configured to switch. Each of the independent relays (109) is local AI 25 by processor, neural network module, relay control and Bluetooth module (108) It is switched on or off according to the control signal generated. This structure Thanks to this, there is no need to cut off the power to the entire system; only the power can be cut off. devices that are consuming unnecessary energy, are in standby mode, or pose a risk The power outlet it is connected to is being disabled. The decision to cut off the power depends on whether the connected device is in active use or in standby mode. how long it stays on, user habits, dependency between different power outlets 14 The relationship and anomaly situation are evaluated together to determine the user's status. When it is detected that the device is still being used, the power supply to the relevant socket is turned off. This continues. Prolonged waiting periods outside of the device's defined usage pattern are avoided. if it remains in mode or the use of the other device it is connected to ends The power supply is interrupted by controlling the independent relay (109). 5 System status information, connection information, and fire risk information are controlled via WiFi / Bluetooth. is displayed on the panel (102). WiFi / Bluetooth control panel (102), depending on whether the system is active or passive, the wireless communication status, and the power outlet. Information regarding the working status and identified risk level of the entries 10 It presents to the user. The said control panel (102) presents the system locally. It enables monitoring and the transmission of necessary user commands to the system. Local AI processor, neural network module, relay control and Bluetooth module (108) The Bluetooth communication infrastructure provides a local 15-inch connection between the system and the user device. It enables wireless communication. If a WiFi connection is available. System information WiFi / Bluetooth control panel (102) and related communication components It is transferred to the mobile application via [internet connection]. [No internet connection required] In this case, device recognition, habit learning, addiction analysis, anomaly detection, and Relay control operations are performed locally. 20 The offline mode indicator (115) indicates that the system is operating independently of the internet connection. It is communicated to the user via the Offline mode indicator (115), the system's external that it performs local data processing without a server or cloud connection It shows that during offline operation, the CT current sensors (106) and temperature are 25. Data from the flame sensor array (110) continues to be processed, local AI and The anomaly detection engine (113) continues to assess the risk and, if necessary Independent relays (109) are controlled. Mobile application notification output (114), system generated status, energy 30 It enables the transmission of cutting and risk information to the user's device. Mobile application As a result of device identification via notification output (114), the socket inputs are active or passive. the situation, the automated power cut-off process, anomaly detection and fire risk The warning is being conveyed to the user. The user can check the operating status of the power outlets (101) via the mobile application. 5 displays and monitors the control processes performed by the system. The system allows the user to use a specific power outlet in the permitted operating mode. manual for re-energizing or de-energizing the input It sends a command. The command sent is the mobile application notification output (114), WiFi / Bluetooth control panel (102) and local AI processor, neural network module, relay control and via Bluetooth module (108) to the relevant independent relay (109) 10 is being transferred. A learning process takes place after the initial system setup. Learning Power consumption data for each socket input (101) during the process are collected by CT current sensors (106) This data is detected by the power signature analog-to-digital converter and 15 The frequency is digitized by the frequency analysis unit (107) and the frequency is digitized by the FFT algorithm. It is decomposed into its components. The resulting power signatures are used to create the device power signature graph (111) It is represented on and includes a local AI processor, neural network module, relay control, and It is processed by the Bluetooth module (108). In the learning process, device usage hours, active working hours, and standby hours are considered. Their status and usage relationships with each other are recorded. Different sockets. Data relating to inputs within the scope of the input dependency analysis scheme (112) They are compared and a user-specific usage model is created. Learning As the process progresses, device recognition, usage habit determination, and power cut-off 25 The accuracy of their decisions is being updated. Examples of use include a laptop charger, a game console, and a desk lamp. CT current sensors (106) are connected to different socket inputs (101) of each device It detects the current it draws separately and the power signature is determined by an analog-to-digital converter with a frequency of 30. The analysis unit (107) generates the digital power signatures of the devices. The generated power The signatures are processed via the device power signature graph (111) and the local AI processor, neural 16 Network module, relay control and Bluetooth module (108) with device profiles They are being matched. After the laptop's charging process is complete, power consumption is low and It has been determined that it has reached a stable level. This situation is based on the previously learned charge level of 5. If it matches the completion pattern, the local AI processor, neural network module, and relay control and Bluetooth module (108), independent relay (109) of the relevant socket input It cuts off the power supply by controlling it. Using the game console with a television or display device 10 In this case, the input dependency analysis diagram of the usage data of the two devices (112) It is associated with it. The game continues after the display device is turned off. If it is determined that the console remains in sleep mode, the learned dependency According to the relationship, the power supply to the game console's power outlet is provided by the relevant independent relay. (109) is cut off through. 15 If the system learns that the desk lamp is only used during certain hours, It records these usage times in a user habit model. Table If the lamp is activated outside of its usual operating hours, the system taking into account the real-time usage of the device and the device being actively used 20 It maintains the power supply as long as the device is in use. When the device is not in use or A decision to cut off the power supply will be made if it is determined that the system has been in a standby state for a long period of time. is being created. If any temperature increase occurs at any socket input (101), the temperature and 25 The flame sensor array (110) detects the relevant data. Simultaneously, from the CT current sensor (106) In case of high or unusual current consumption, the data in question The local AI and anomaly detection engine (113) are evaluated together. The risk level is displayed on the fire risk score bar indicator (103), the relevant independent Power is cut off via relay (109) and mobile application notification output (114) 30 The alert is transmitted to the user's device via this method. 17 Energy monitoring as long as the system is powered via the main power cable connection (105), device recognition, habit learning, inter-input dependency analysis, temperature and flame. Detection, anomaly detection, fire risk score generation, and independent relay control. It performs these operations repeatedly. As usage data is updated, the device power... Signatures, user habits, and dependency relationships between inputs are re-evaluated. 5 is being evaluated. This structure includes the main body (100), power outlets (101), and WiFi / Bluetooth control panel. (102), fire risk score bar indicator (103), USB-A and USB-C output ports (104), main power cable connection (105), CT current sensors (106), power signature analog-digital 10 converter and frequency analysis unit (107), local AI processor, neural network module, relay control and Bluetooth module (108), independent relays (109), temperature and flame sensor array (110), device power signature graph (111), input dependency analysis diagram (112), local AI and anomaly detection engine (113), mobile application notification output (114) and The offline mode indicator (115) works in a functional connection with each other. 15 25

Claims

18 REQUESTS 1- The invention relates to a smart socket with multiple inputs and a learning energy management system, the feature of which is;  a main body (100),  Multiple socket inputs located on the main body (100) 5 (101),  CT current sensors detecting power consumption data for each socket input (101) (106),  Analog power consumption data obtained from CT current sensors (106) power signature analog-to-digital converter and 10 that create power signature by digitizing it Frequency analysis unit (107),  By processing the generated power signature, the user identifies the connected device. learning their habits, performing dependency analysis between inputs, and The local AI processor, neural network module, and relay that make the decision to cut off the power. control and Bluetooth module (108), 15  in line with the control signals generated by the local AI processor (108) independent switches that independently switch the power supply to each socket outlet. relay (109),  Temperature and flame sensor that detects temperature and flame information for each socket input. sensor array (110), 20  Creates a fire risk score by processing temperature, current, and anomaly data. fire risk score bar indicator (103),  WiFi / Bluetooth displaying system status information and fire risk information. control panel (102),  Mobile application notification that transmits status and warning information to the user's device 25 It consists of (111) output. 2- The smart plug mentioned in Claim 1 is a multi-input learning energy management system, Its feature is a power signature that separates power consumption data into frequency components using an FFT algorithm. 30 characterized by having an analog-to-digital converter and a frequency analysis unit (107). It is done. 19 3- The smart plug mentioned in Claim 1 is a multi-input learning energy management system, This feature identifies the connected electronic device by storing power signature data in memory. It has a local AI processor, neural network module, relay control and Bluetooth module (108). It is characterized by its being. 4- The smart plug mentioned in Claim 1 is a multi-input learning energy management system, This feature processes time-based usage data belonging to the user. a local AI processor that determines its habits, a neural network module, relay control, and It is characterized by having a Bluetooth module (108). 5- The smart plug mentioned in Claim 1 is a multi-input learning energy management system, Its feature analyzes the time relationship between usage data from different power outlets. a local AI processor, neural network module, and relay that determine the dependency between inputs. It is characterized by having a control and Bluetooth module (108). 6- The smart plug mentioned in Claim 1 is a multi-input learning energy management system, Its feature is to detect anomalies by evaluating temperature data and current consumption data together. It is characterized by having a temperature and flame sensor array (110) It is done. 7- The smart plug mentioned in Claim 1 is a multi-input learning energy management system, The feature is that each one is based on the control signals generated by the local AI processor (108). It is characterized by having independent relays (109) that switch the socket input separately. It is done. 8- The smart plug mentioned in Claim 1 is a multi-input learning energy management system, Its feature is that it transmits system status information and fire risk information via wireless communication. mobile application has notification output (111) which transmits to the user device It is the characterization of the situation. 9- The invention is the method for a learning energy management system with multiple input smart sockets. feature;  CT current sensors (106) of power consumption data for each socket input perception through,  Power signature of perceived power consumption data analog-to-digital converter and frequency digitization by the analysis unit (107) and frequency components separation, 5  the generated power signature data is processed by the local AI processor (108) Identifying the connected device,  By processing user usage data, usage habits are analyzed. determination,  The dependency relationship between usage data for different power outlets is 10 determination,  Generating anomaly information by evaluating temperature and current data,  Control of independent relays (109) according to the analysis results created,  Mobile application notification output of generated system status and risk information. (111) transmitted to the user device 15 It is a method characterized by including the following steps. 10- This is the method mentioned in claim 9, and its characteristic is that it uses an FFT algorithm to process power consumption data. It is characterized by including the step of decomposing it into its frequency components using It is done. 20 11- This is the method mentioned in Claim 9, and its characteristic is that it uses the generated power signature. It is characterized by including the step of identifying the connected electronic device. 12- The method mentioned in Request 9, its characteristic is; usage of different socket inputs 25 This includes the step of determining the dependency relationship between inputs using the data. It is the characterization of the situation. 13- This is the method mentioned in Claim 9, and its characteristic is that it combines temperature and current data. by evaluating the anomaly, information about the anomaly is generated and based on this information, 30 It is characterized by the inclusion of the step of controlling independent relays (109).