Methods, devices, and systems for predicting the condition of apartment complex facilities based on artificial intelligence and IoT sensors, and for integrated management of complex-specific accounting and material linkage operations

Abstract

The device according to one embodiment collects data by apartment complex based on an apartment complex database, groups it according to preset criteria, collects equipment status data through IoT sensors installed on each facility within the grouped apartment complex, inputs the equipment status data into an artificial intelligence model pre-trained based on past equipment failure history data and inspection result data to generate equipment status prediction information including the probability of abnormality occurring in each facility and the timing of maintenance required, generates ordering information for materials required within a preset period by referencing inventory information by the grouped complex stored in a material management database based on the equipment status prediction information, adjusts the operational plan by the grouped complex, including security schedules, cleaning schedules, and common area management schedules, according to preset criteria based on the equipment status prediction information, automatically generates accounting management information by the grouped complex, including maintenance costs, labor costs, material costs, and contingency reserves based on the equipment status prediction information and the operational plan by the grouped complex, and can integrate, store, and manage management data by the grouped complex including equipment status prediction, material ordering, operational plan adjustment, and accounting management.

Description

Technology Field

[0001] The following embodiments relate to technology that provides a method, device, and system for predicting the status of multi-unit housing facilities based on artificial intelligence and IoT sensors, and for integrated management of accounting and material linkage operations by complex. Background Technology

[0003] Multi-unit housing complexes are living spaces where multiple households share the same infrastructure, consisting of various facilities such as electricity, water, elevators, fire safety systems, lighting, and ventilation. These facilities must be inspected and maintained at regular intervals, and any breakdowns or deterioration directly affect the safety and convenience of residents.

[0004] Existing facility management in multi-unit residential buildings largely relies on manual inspection methods centered on manpower. Managers often only become aware of problems and request repairs or replacements from outsourced companies after equipment failures or complaints occur. Consequently, problems frequently arose where signs of equipment malfunctions were not detected in advance, leading to delayed responses after breakdowns occurred. Furthermore, inspection cycles and items were set uniformly without considering the specific characteristics of each piece of equipment, resulting in unnecessary inspections and wasted budget.

[0005] With the advancement of IoT technology, techniques for collecting real-time status data such as equipment temperature, current, vibration, and pressure have emerged; however, there are few instances where this sensor data is effectively utilized in actual management systems. Most systems operate at a level limited to simple monitoring or detecting only abnormalities in specific equipment. In other words, there is a limitation in that equipment-specific data is not linked to overall management decision-making (e.g., budgeting, material purchasing, personnel allocation, etc.).

[0006] Furthermore, accounting and materials management, which are other core areas of apartment complex management, are also generally operated separately within individual systems. For instance, maintenance budgets often fail to reflect the actual condition of facilities or replacement cycles, instead relying on past expenditure records or the manager's experience. This leads to budget overruns and unnecessary spending, while material ordering also results in issues such as preemptive or duplicate orders placed regardless of actual timing of need.

[0007] In particular, despite the fact that the scale, age of facilities, and characteristics of residents vary by complex, most management systems are operated under uniform standards. As a result, maintenance priorities and the timing of material procurement are not properly coordinated, leading to problems where management efficiency differs between complexes.

[0008] Therefore, in the field of multi-unit housing management, there is a demand for technology capable of supporting operational decision-making tailored to the specific characteristics of each complex by comprehensively analyzing facility status data with accounting and material information. In other words, a new integrated management technology is required that utilizes artificial intelligence and IoT sensor technology to predict facility status and links the results to complex-specific accounting and material management, thereby simultaneously improving operational efficiency and the transparency of budget execution.

[0009] Therefore, technology is required to provide methods, devices, and systems for predicting the status of multi-unit housing facilities based on artificial intelligence and IoT sensors, as well as for integrated management of accounting and material-linked operations for each complex. Prior art literature

[0011] Republic of Korea Registered Patent No. 10-2380397 (Published March 31, 2022) Republic of Korea Registered Patent No. 10-2464762 (Published November 9, 2022) Republic of Korea Registered Patent No. 10-1823544 (Published January 31, 2018) Republic of Korea Registered Patent No. 10-2690705 (Published August 5, 2024) The problem to be solved

[0012] The embodiments aim to provide a method for predicting the facility status of a multi-unit housing complex using artificial intelligence and IoT sensors, and for integrated management by linking accounting and material information.

[0013] The embodiments aim to provide a method for grouping according to complex characteristics and automatically calculating material ordering and accounting items based on equipment forecast results.

[0014] The embodiments aim to provide a method for establishing intensive patrol zones by analyzing zone-specific risk indices and generating a patrol plan that reflects real-time data.

[0015] The embodiments aim to provide a method for predicting the lifecycle and automatically generating replacement and maintenance schedules by analyzing equipment operation data and environmental information.

[0016] The embodiments aim to provide a method for analyzing management data by complex to evaluate operational efficiency and cost reduction rates and generating a performance report.

[0017] The objectives of the present invention are not limited to those mentioned above, and other unmentioned objectives will be clearly understood from the description below. means of solving the problem

[0019] According to one embodiment, in a method for predicting the condition of multi-unit housing facilities based on artificial intelligence and IoT sensors and for integrated management of accounting and material-linked operations by complex, the method comprises: a step of collecting data by multi-unit housing complex based on a multi-unit housing database and grouping it according to preset criteria; a step of collecting facility status data through IoT sensors installed on each facility within the grouped multi-unit housing complex; a step of generating facility status prediction information including the probability of abnormality occurring in each facility and the timing of maintenance requirements by inputting the facility status data into an artificial intelligence model that has been pre-trained based on past facility failure history data and inspection result data; a step of generating ordering information for materials required within a preset period by referring to the grouped complex-specific inventory information stored in a material management database based on the facility status prediction information; a step of adjusting the grouped complex-specific operation plan, including security schedules, cleaning schedules, and common area management schedules, according to preset criteria based on the facility status prediction information; and a step of automatically generating grouped complex-specific accounting management information including maintenance costs, labor costs, material costs, and contingency reserves based on the facility status prediction information and the grouped complex-specific operation plan. and may include the step of integrating, storing, and managing grouped complex-specific management data, including facility status prediction, material ordering, operation plan coordination, and accounting management.

[0020] The step of grouping the above-mentioned multi-unit housing complexes according to pre-set criteria may include: creating a first group for maintenance efficiency based on the degree of facility obsolescence, facility type, and number of households for each multi-unit housing complex; creating a second group for management cost reduction based on energy usage and maintenance cost patterns for each multi-unit housing complex; creating a third group for service operation efficiency based on the age group, income level, and lifestyle patterns of residents for each multi-unit housing complex; and setting data processing weights corresponding to the management purpose for each group based on the first group, the second group, and the third group.

[0021] The step of generating equipment condition prediction information including the probability of anomaly occurring in each of the above-mentioned facilities and the timing of maintenance required may include: a step of extracting change patterns of current values, temperature values, pressure values, and vibration values ​​from the equipment condition data; a step of inputting the extracted change patterns of the equipment condition data into the artificial intelligence model and calculating an anomaly indication score according to a preset standard by comparing it with a past normal state pattern; a step of calculating the probability of anomaly occurring according to a preset standard by combining the calculated anomaly indication score with the aging index, usage frequency, and recent inspection cycle for each facility; a step of predicting the timing of maintenance required and the scope of maintenance work by applying a preset equipment importance weight to the calculated probability of anomaly occurring; and a step of generating equipment condition prediction information including the probability of anomaly occurring, the timing of maintenance required, and the scope of maintenance work, and reflecting the generated equipment condition prediction information in the grouped management data for each complex.

[0022] The step of adjusting the operation plan for each grouped complex may include: a step of calculating the risk of each facility based on the predicted probability of facility malfunction and the timing of maintenance requirements; a step of setting priorities for security, cleaning, and common area management tasks based on the calculated risk; a step of adjusting the placement of security personnel, cleaning personnel, and inspection personnel according to preset criteria based on the set priorities, and automatically generating a work schedule by considering the estimated time required for each task, the number of available personnel, and movement paths; a step of calculating an efficiency index for each task according to preset criteria to evaluate the management efficiency of the generated work schedule; a step of regenerating the work schedule if the efficiency index is less than a preset threshold value; and a step of adjusting the work schedule where the efficiency index is greater than or equal to a preset threshold value by reflecting it in the operation plan for the grouped complex.

[0023] The step of automatically generating accounting management information for each grouped complex may include: a step of confirming the timing of maintenance required and the scope of maintenance work for each facility based on the facility status prediction information, and determining the time required for work by querying a preset work time value corresponding to the work scope; a step of confirming the work schedule for each personnel based on the complex operation plan, and calculating labor costs based on the time required for work and a preset labor cost unit price; a step of calculating maintenance costs based on the preset unit price for each task based on the predicted maintenance work scope; a step of calculating material costs by confirming a preset material unit price based on the generated material ordering information; a step of calculating reserve funds by applying the reserve ratio to the sum of maintenance costs, labor costs, and material costs based on a preset reserve ratio; a step of calculating the total amount of accounting items by summing the maintenance costs, labor costs, material costs, and reserve funds; and a step of determining whether the calculated total amount of accounting items exceeds a preset budget limit by comparing it with the calculated total amount of accounting items, generating budget alert data if it exceeds the limit, and reflecting the total amount of accounting items in the grouped complex management data if it does not exceed the limit.

[0024] The method further includes the steps of: calculating a risk index for each zone within a multi-unit housing complex to establish a concentrated patrol zone and generating a patrol plan; and reflecting the generated patrol plan in the grouped operation plan for each complex; wherein the step of establishing the concentrated patrol zone may include: calculating a potential risk index for each zone by verifying past alarm history, crime frequency, and facility failure history for each zone within the multi-unit housing complex based on the multi-unit housing database; collecting real-time video data from CCTVs installed in the multi-unit housing complex to extract real-time brightness data for each installation zone of lighting equipment, including streetlights and security lights; analyzing the video data collected from the CCTVs to detect abnormal events according to pre-set criteria and generating real-time alarm data; integrating the potential risk index, real-time brightness data for each zone, and real-time alarm data to calculate an integrated patrol risk index for each zone; establishing a zone where the integrated patrol risk index is greater than or equal to a pre-set threshold value as a concentrated patrol zone; and generating a grouped patrol plan for each complex by calculating the number of patrols, patrol routes, and patrol times according to pre-set criteria for the concentrated patrol zone.

[0025] The method further includes the step of predicting the lifecycle of facilities by grouped complex based on the facility status prediction information and generating replacement and maintenance schedules; wherein the step of predicting the lifecycle of facilities and generating replacement and maintenance schedules comprises: a step of calculating the cumulative operating time and average load rate for each facility based on operating time data, load rate data, and environmental variable data collected through IoT sensors installed in each facility within the multi-unit housing complex; a step of calculating the remaining lifespan for each facility according to a preset standard lifespan value for each facility by comparing the cumulative operating time, average load rate, and environmental variable data with a preset standard lifespan value for each facility; a step of generating early warning data for facilities with a high probability of early failure by comparing the remaining lifespan calculation result with past failure history data and manufacturer warranty period information; a step of calculating the failure probability based on failure history data of similar facility types and calculating the facility replacement priority using the failure probability and remaining lifespan data; a step of determining the facility replacement time and maintenance cycle according to a preset standard based on the replacement priority and the remaining lifespan for each facility; a step of collecting performance data through the IoT sensors and calculating a performance deviation index by comparing it with a reference performance value; and the performance deviation index is a preset reference value If exceeded, the method may include a step of automatically recommending a response manual by referring to pre-set failure cause data by equipment type, and a step of calculating a risk level by equipment by integrating information on remaining lifespan, replacement priority, maintenance cycle, performance deviation index, and response manual by equipment, predicting the life cycle of equipment based on the risk level and maintenance cycle, and generating a replacement and maintenance schedule according to the predicted life cycle and risk level.

[0026] The method further includes the step of analyzing operational efficiency among grouped complexes based on the grouped complex-specific management data and generating a performance report based on the analysis results; wherein the step of generating a performance report based on the grouped complex-specific management data comprises: extracting maintenance cost, labor cost, and material cost data from accounting management information for each grouped complex and calculating a cost reduction rate for each complex by comparing it with the average value for each complex within the group; extracting work schedule and personnel allocation data from operational planning information for each grouped complex and calculating a manpower input efficiency index according to preset criteria; verifying material unit price and delivery date data from material ordering information and material management database for each grouped complex and calculating a material procurement efficiency index for each complex by comparing it with the average value within the group; calculating an overall operational efficiency index for the grouped complex by combining the cost reduction rate, manpower input efficiency index, and material procurement efficiency index; determining whether efficiency has improved by comparing the overall operational efficiency index with a preset standard value and generating data requiring improvement if it falls below the standard value; and generating a performance report including the overall operational efficiency index, cost reduction rate, manpower input efficiency index, and material procurement efficiency index, and the generated performance report for multi-unit housing It may include a step of transmitting to the administrator's terminal.

[0027] A device according to one embodiment may be combined with hardware and controlled by a computer program stored on a medium to execute the method of any one of the methods described above. Effects of the invention

[0029] The embodiments can provide a method to improve the efficiency and stability of complex operations by utilizing artificial intelligence and IoT sensors to predict the facility status of a multi-unit housing complex and by integrating and managing accounting and material information.

[0030] The embodiments can provide a method to reduce management costs and increase the accuracy of budget management by automatically calculating material ordering and accounting items based on grouping reflecting complex-specific characteristics and equipment forecasting results.

[0031] The embodiments can provide a method to improve the level of safety management within the complex by utilizing zone-specific risk indices and real-time data to establish intensive patrol zones and generate efficient patrol plans.

[0032] The embodiments can provide a method to improve the stable operation and maintenance efficiency of equipment by analyzing equipment operation data and environmental information to predict the lifecycle and automatically generate replacement and maintenance schedules.

[0033] The embodiments can provide a method to objectively compare management performance between complexes and derive directions for improvement by analyzing management data by complex to evaluate operational efficiency and cost reduction rates and generating performance reports.

[0034] Meanwhile, the effects according to the embodiments are not limited to those mentioned above, and other unmentioned effects will be clearly understood by those skilled in the art from the description below. Brief explanation of the drawing

[0036] FIG. 1 is a schematic diagram showing the configuration of a system according to one embodiment. FIG. 2 is a flowchart illustrating the process of predicting the condition of multi-unit housing facilities and managing integrated management of accounting and material linkage operations by complex according to one embodiment. FIG. 3 is a flowchart illustrating the process of grouping a multi-unit housing complex according to one embodiment. FIG. 4 is a flowchart illustrating the process of generating equipment status prediction information according to one embodiment. FIG. 5 is a flowchart illustrating the process of coordinating operation plans for each grouped complex according to one embodiment. FIG. 6 is a flowchart illustrating the process of automatically generating accounting management information for grouped complexes according to one embodiment. FIG. 7 is a flowchart illustrating the process of establishing a concentrated patrol zone according to one embodiment. FIG. 8 is a flowchart illustrating the process of predicting the life cycle of equipment and generating replacement and maintenance schedules according to one embodiment. FIG. 9 is a flowchart illustrating the process of generating a performance report based on grouped management data by complex according to one embodiment. FIG. 10 is an example diagram of the configuration of a device according to one embodiment. Specific details for implementing the invention

[0037] Hereinafter, embodiments are described in detail with reference to the attached drawings. However, various modifications may be made to the embodiments, and thus the scope of the patent application is not limited or restricted by these embodiments. It should be understood that all modifications, equivalents, and substitutions to the embodiments are included within the scope of the rights.

[0038] Specific structural or functional descriptions of the embodiments are disclosed for illustrative purposes only and may be modified and implemented in various forms. Accordingly, the embodiments are not limited to the specific disclosed forms, and the scope of this specification includes modifications, equivalents, or substitutions that fall within the technical concept.

[0039] Terms such as "first" or "second" may be used to describe various components, but these terms should be interpreted solely for the purpose of distinguishing one component from another. For example, the first component may be named the second component, and similarly, the second component may be named the first component.

[0040] When it is stated that a component is "connected" to another component, it should be understood that it may be directly connected to or joined to that other component, or that there may be other components in between.

[0041] The terms used in the embodiments are for illustrative purposes only and should not be interpreted as intended to be limiting. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this specification, terms such as "comprising" or "having" are intended to indicate the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0042] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which the embodiments pertain. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this application.

[0043] In addition, when describing with reference to the attached drawings, identical components are assigned the same reference numeral regardless of drawing symbols, and redundant descriptions thereof are omitted. When describing embodiments, if it is determined that a detailed description of related prior art could unnecessarily obscure the essence of the embodiment, such detailed description is omitted.

[0044] The embodiments can be implemented in various forms of products, such as personal computers, laptop computers, tablet computers, smartphones, smart home appliances, intelligent automobiles, kiosks, and wearable devices.

[0045] FIG. 1 is a schematic diagram showing the configuration of a system according to one embodiment.

[0046] Referring to FIG. 1, a system according to one embodiment may include a terminal (100) and a device (200) of a multi-unit housing manager that can communicate with each other through a communication network.

[0047] First, the communication network can be configured regardless of the mode of communication, such as wired or wireless, and can be implemented in various forms to enable communication between servers and between servers and terminals.

[0048] The terminal (100) of the apartment complex manager is a terminal used by the manager to comprehensively view and control the status of facilities, material management, accounting information, and operation plans of the apartment complex.

[0049] The terminal (100) of the apartment building manager is connected to the device (200) via a network, and can receive equipment status prediction information, material ordering information, accounting management information, and patrol plan information generated by the device (200) in real time.

[0050] That is, the terminal (100) of the apartment complex manager can manage multiple apartment complexes as a single management group, rather than having an independent management structure for each complex as in the past, and can manage them in an integrated manner on a group basis.

[0051] For example, the terminal (100) of a multi-unit housing manager can set 30 complexes within “A Region” as one group and integrate and compare data on the facility status, inspection cycle, budget execution status, personnel deployment, and material inventory of each complex on one screen.

[0052] This allows apartment complex managers to coordinate maintenance plans on a group basis rather than on an individual complex basis, and to flexibly reallocate personnel or materials where needed.

[0053] In addition, the terminal (100) of the apartment building manager is configured to indirectly receive equipment status data such as current, pressure, vibration, temperature, and humidity from IoT sensors installed in each complex, thereby enabling the detection of abnormal signs in real time.

[0054] For example, if a pressure sensor installed in the piping of a specific complex detects a blockage, an alarm generated by the device (200) is displayed on the terminal (100) of the apartment complex manager, and the apartment complex manager can immediately issue an inspection order to the patrol personnel of the nearby complex.

[0055] At this time, the terminal (100) of the apartment building manager can display the current location and work schedule of security, cleaning, and inspection personnel together, thereby enabling efficient on-site response management.

[0056] The terminal (100) of the apartment building manager may also include material management and automatic ordering functions.

[0057] When consumables or materials are used in each complex and the inventory quantity decreases below a preset standard, the terminal (100) of the apartment complex manager can be linked with the device (200) to generate an automatic order request for the item.

[0058] For example, when only 2 out of 10 lamps for light replacement remain, the terminal (100) of the apartment building manager can display a “lamp stock shortage” notification and automatically place an order with the supplier through the device (200).

[0059] In this way, the terminal (100) of the apartment building manager can manage material inventory, order details, and delivery status by complex in real time.

[0060] The terminal (100) of the apartment building manager includes an accounting management automation function, automatically sums up items such as maintenance costs, labor costs, material costs and reserve costs, and can display the results of calculating management fees for each complex based on accounting management information received from the device (200).

[0061] For example, the terminal (100) of the apartment building manager can display whether the budget has been exceeded by comparing the total amount of accounting items generated by the device (200) with a preset budget limit, or visualize the expenditure details (maintenance costs, material costs, labor costs, etc.) by accounting item as an item-by-item graph.

[0062] In addition, the terminal (100) of the apartment building manager can perform resident notification and automatic billing by linking the calculated management fee data with the resident management system.

[0063] The terminal (100) of the apartment building manager may also include a function for analyzing operational efficiency by grouped complexes.

[0064] For example, the terminal (100) can compare and visually display the cost reduction rate, manpower input efficiency, and material procurement efficiency of each complex within the group, and can also present data on the need for improvement received from the device (200) for complexes with low efficiency.

[0065] Through this, managers can intuitively grasp operational efficiency by complex and immediately perform resource reallocation or operational plan adjustments.

[0066] In addition, the terminal (100) of the apartment building manager can provide patrol plan monitoring and control functions.

[0067] For example, if a decrease in illumination or abnormal activity in a designated area is detected through real-time video data received from a CCTV or drone patrol system, the manager can directly readjust the patrol cycle or route of the area on the terminal screen.

[0068] At this time, the terminal (100) of the apartment building manager can display the integrated patrol risk indicator generated by the device (200) together, thereby supporting the manager in performing efficient patrol management based on priority.

[0069] Through this, the terminal (100) of the apartment building manager can grasp the management efficiency of each complex and group at a glance, track the status of IoT-based facilities and material consumption in real time, and automatically perform material ordering, accounting settlement, and patrol control in conjunction with the device (200), thereby enabling intelligent integrated management that replaces the existing manual management structure for each complex.

[0070] The terminal (100) of the apartment building manager may be a desktop computer, laptop, tablet, smartphone, etc. For example, the terminal (100) of the apartment building manager may be a tablet, and may be adopted differently depending on the embodiment.

[0071] The terminal (100) of the apartment building manager may be configured to perform all or part of the computational function, storage / reference function, input / output function and control function that a normal computer has, and the terminal (100) of the apartment building manager may be configured to communicate with the device (200) via wired or wireless communication.

[0072] The terminal (100) of the apartment building manager may be connected to a website operated by a person or organization providing services using the device (200), or may have an application developed and distributed by a person or organization providing services using the device (200) installed. The terminal (100) of the apartment building manager may be linked with the device (200) through the website or application.

[0073] The terminal (100) of the apartment building manager can access the device (200) through a web page, application, etc. provided by the device (200).

[0074] The device (200) may be a private server owned by a person or organization providing a service using the device (200), a cloud server, or a peer-to-peer (P2P) set of distributed nodes. The device (200) may be configured to perform all or part of the computational functions, storage / reference functions, input / output functions, and control functions that a conventional computer possesses.

[0075] The device (200) can be configured to communicate wirelessly or via wired connection with the terminal (100) of the apartment building manager.

[0076] The device (200) can predict the facility status of a multi-unit housing complex using artificial intelligence and IoT sensors, and integrate management of operations by complex by linking accounting and material information.

[0077] The device (200) can group apartment complexes according to the degree of equipment deterioration, energy consumption, and resident characteristics, and can automatically calculate material ordering and accounting items based on the results of equipment condition prediction.

[0078] The device (200) can analyze the risk index by zone and automatically generate a concentrated patrol zone and patrol plan by reflecting real-time video and sensor data.

[0079] The device (200) can analyze the operation data, load rate, and environmental information of the equipment to predict the life cycle of the equipment and automatically generate replacement and maintenance schedules.

[0080] The device (200) can analyze grouped management data by complex to evaluate cost reduction rate, manpower input efficiency and material procurement efficiency, and generate a performance report by complex based on this.

[0081] Through this, the device (200) can predict the facility status of a multi-unit housing complex by utilizing artificial intelligence and IoT sensors, and by linking the prediction results to material management, accounting management, operation planning, and safety management, it can simultaneously improve efficient operation of each complex, transparency of budget execution, and stability of facility management.

[0082] The process of predicting the status of multi-unit housing facilities based on artificial intelligence and IoT sensors, as well as the integrated management of accounting and material linkage operations for each complex, can be carried out through a server that includes a processor for collecting and processing such information. The service may be provided via a web-based platform or a smartphone application, and in some cases, processing can be performed by applying artificial neural networks or machine learning.

[0083] Additionally, the device (200) can communicate wirelessly or via wired connection with websites including social media platforms such as blogs, cafes, Instagram, Facebook, Twitter, and YouTube, and web pages including articles, and the device (200) can access the websites to obtain information.

[0084] Meanwhile, for convenience of explanation, only one terminal (100) of a multi-unit housing manager is shown in FIG. 1 and the following description, but the number of terminals can vary depending on the embodiment. As long as the processing capacity of the device (200) allows, there is no particular limit to the number of terminals.

[0085] In the present invention, Artificial Intelligence (AI) refers to a technology that imitates human learning ability, reasoning ability, and perceptual ability, and implements them on a computer, and may include concepts such as machine learning and symbolic logic. Machine Learning (ML) is an algorithmic technology that classifies or learns the characteristics of input data on its own. AI technology can analyze input data as a machine learning algorithm, learn from the results of the analysis, and make judgments or predictions based on the results of the learning. Furthermore, technologies that mimic the functions of the human brain, such as cognition and judgment, by utilizing machine learning algorithms can also be understood as falling within the category of AI. For example, technological fields such as linguistic understanding, visual understanding, reasoning / prediction, knowledge representation, and motion control may be included.

[0086] Machine learning can refer to the process of training neural network models using experience in processing data. It implies that through machine learning, computer software improves its own data processing capabilities. A neural network model is constructed by modeling the correlations between data, and these correlations can be expressed by multiple parameters. A neural network model extracts and analyzes features from given data to derive correlations between them; machine learning can be defined as the process of optimizing the model's parameters by repeating this process. For example, a neural network model can learn the mapping (correlation) between inputs and outputs for data given as input-output pairs. Alternatively, even when only input data is provided, a neural network model can derive regularities between the given data and learn those relationships.

[0087] An artificial intelligence learning model or neural network model can be designed to implement the structure of the human brain on a computer and may include multiple network nodes that have weights and simulate neurons of a human neural network. The multiple network nodes may have interconnected relationships by simulating the synaptic activity of neurons, where neurons exchange signals through synapses. In an artificial intelligence learning model, multiple network nodes may be located in layers of different depths and exchange data according to convolutional connections. The artificial intelligence learning model may be, for example, an Artificial Neural Network (ANN) or a Convolutional Neural Network (CNN). As an embodiment, the artificial intelligence learning model may be machine learned according to methods such as supervised learning, unsupervised learning, and reinforcement learning. Machine learning algorithms for performing machine learning may include Decision Tree, Bayesian Network, Support Vector Machine, Artificial Neural Network, Ada-boost, Perceptron, Genetic Programming, and Clustering.

[0088] Among these, CNNs are a type of multilayer perceptron designed to use minimal preprocessing. CNNs consist of one or more convolutional layers and standard artificial neural network layers stacked on top, additionally utilizing weights and pooling layers. Thanks to this structure, CNNs can fully utilize two-dimensional input data. Compared to other deep learning architectures, CNNs demonstrate good performance in both image and audio fields. CNNs can also be trained using standard backpropagation. CNNs have the advantage of being easier to train than other feedforward artificial neural network techniques and using a small number of parameters.

[0089] Convolutional networks are neural networks comprising sets of nodes with bounded parameters. Many computer vision tasks have been significantly improved, driven by the increased size of available training data and the availability of computational power, combined with algorithmic advancements such as discriminative linear units and dropout training. In the case of massive datasets, such as those available for many tasks today, outfitting is not critical, and increasing the network size improves test accuracy. Optimal utilization of computing resources becomes a limiting factor. To address this, distributed, scalable implementations of deep neural networks can be employed.

[0090] FIG. 2 is a flowchart illustrating the process of predicting the condition of multi-unit housing facilities and managing integrated management of accounting and material linkage operations by complex according to one embodiment.

[0091] Referring to FIG. 2, first, in step S201, the device (200) can collect data by apartment complex based on an apartment complex database and group it according to preset criteria.

[0092] That is, the device (200) can analyze management data by complex stored in a multi-unit housing database and perform grouping by complex based on the degree of facility deterioration, facility type, number of households, energy consumption, maintenance costs, and living characteristics of residents of each complex.

[0093] Here, the multi-unit housing database refers to a database containing the facility management history, energy usage, regular inspection results, personnel deployment details, and accounting expenditure data of each complex, and the device (200) can use this to construct a normalized dataset that allows comparison between complexes.

[0094] For example, the device (200) can classify complexes with an average facility age of 10 years or more into a first group requiring maintenance efficiency, complexes with high maintenance costs relative to energy consumption into a second group requiring management cost reduction, and complexes with high age groups and income levels of residents into a third group requiring service operation efficiency.

[0095] At this time, the pre-set criteria may be set to items such as an equipment obsolescence index of 0.7 or higher, an annual energy consumption per household of 15% or higher compared to the average of the complex, and an energy efficiency index of 0.8 or lower relative to maintenance costs, etc., but this is merely an example and may be set differently depending on the embodiment.

[0096] Additionally, the device (200) can set data processing weights according to the management purpose of each group. For example, a weight of 0.6 for 'equipment condition prediction accuracy' can be applied to the first group, a weight of 0.5 for 'cost reduction effect' to the second group, and a weight of 0.4 for 'service satisfaction' to the third group, and these weight values ​​can be adjusted according to the importance, scale, and management goals of the complex.

[0097] Through this, the device (200) forms a grouping structure that reflects different management objectives for each complex, and subsequently, during the equipment status prediction and operation plan adjustment stage, it can perform optimized analysis and decision-making for each group.

[0098] For a detailed explanation regarding this, refer to Fig. 3.

[0099] In step S202, the device (200) can collect facility status data through IoT sensors installed in each facility within the grouped apartment complex.

[0100] That is, the device (200) can collect real-time data from IoT sensors installed in major facilities within each grouped apartment complex and manage the data in conjunction with the operation history of each facility.

[0101] The device (200) can normalize sensor data into a standardized format and automatically correct missing values ​​or abnormal values ​​to generate highly reliable status information.

[0102] In addition, the device (200) can verify the consistency of measurement values ​​between multiple sensors within the same facility by matching the collected facility status data with the sensor unique identifier (ID) and installation location information.

[0103] For example, if the correlation coefficient between the output values ​​of a pressure sensor and a flow sensor attached to the same pipe is calculated to be less than a preset reference value (e.g., 0.85 or higher), the device (200) may classify the sensor as a 'sensor requiring calibration' and add it to the inspection target list.

[0104] At this time, the correlation coefficient reference value can be set between 0.8 and 0.9 depending on the type of sensor, installation environment, measurement cycle, etc., and this is merely an example and may be set differently depending on the embodiment.

[0105] For example, the device (200) can periodically collect the temperature, pressure, and fuel consumption of the boiler, the vibration value and current usage of the elevator, and the rate of change of pressure of the fire pump, and display data that deviates from a preset normal range (e.g., temperature ±3℃, vibration intensity ±10%) as an abnormal signal.

[0106] At this time, the pre-set standard may be established by referring to the allowable operating range presented by the equipment manufacturer, the average value of past normal operating data, or legal safety standards, etc., and this is merely an example and may be set differently depending on the embodiment.

[0107] Through this, the device (200) can monitor the status of all major facilities within the complex in real time and secure a basis for early detection of abnormal signs.

[0108] In step S203, the device (200) can generate equipment status prediction information including the possibility of anomalies and the time when maintenance is required for each piece of equipment by inputting equipment status data into an artificial intelligence model that has been pre-trained based on past equipment failure history data and inspection result data.

[0109] That is, the device (200) can compare and analyze the collected equipment status data with past failure history data and inspection result data to predict the likelihood of an abnormality occurring for each piece of equipment and the time when maintenance is required.

[0110] The device (200) can distinguish between normal state patterns and abnormal state patterns through a pre-trained artificial intelligence model and convert the risk level of each facility into numerical prediction information.

[0111] Additionally, the device (200) can learn by classifying the state data immediately prior to the occurrence of past equipment failure as an 'abnormal dataset' and the data during the normal operation period as a 'normal dataset' during the learning process of the artificial intelligence model.

[0112] The device (200) can learn the pattern of change in the state of the equipment over time by applying a time series analysis algorithm (e.g., LSTM, GRU) or an anomaly detection algorithm (e.g., Isolation Forest, Autoencoder) and calculate the probability of anomaly occurrence in real time.

[0113] At this time, the training data includes operating time, load rate, temperature and pressure fluctuations, inspection history, etc. for each facility, and the model can be trained repeatedly until the prediction error satisfies a preset standard (e.g., MAPE 5% or less, F1-score 0.8 or more).

[0114] These reference values ​​may vary depending on the type of equipment, failure sensitivity, data collection cycle, etc., and are merely examples and may be set differently depending on the embodiment.

[0115] For example, the device (200) can calculate an abnormal sign score as a probability value between 0 and 1 when the temperature rise rate of the boiler is 5°C or more per hour or the vibration frequency fluctuation rate exceeds 15%, and can determine that the time for maintenance is imminent when this score is 0.7 or higher.

[0116] At this time, the pre-set criteria may be set to include inspection cycles, equipment importance, failure history frequency, manufacturer's recommended maintenance cycles, etc., and this is merely an example and may be set differently depending on the embodiment.

[0117] Through this, the device (200) can establish a data-based decision-making system capable of generating an early warning before equipment failure and establishing a pre-maintenance plan.

[0118] For a detailed explanation regarding this, please refer to Fig. 4.

[0119] In step S204, the device (200) can generate order information for materials required within a preset period by referring to grouped inventory information by complex stored in a material management database based on equipment status prediction information.

[0120] That is, the device (200) can identify parts and consumables of equipment requiring maintenance based on predicted equipment status information, and generate ordering information for necessary materials by referring to a material management database.

[0121] The device (200) can determine the timing of material procurement and order priority by comprehensively reviewing the material-specific inventory quantity, item code, order history, supplier information, and delivery date data stored in the material management database.

[0122] Additionally, the material management database may include data on warehouse locations by complex, material inflow and outflow history, and usage frequency, and the device (200) can calculate the material inventory turnover rate and consumption rate based on this.

[0123] In addition, the device (200) can determine whether to place an order by comparing the inventory level of each material with the 'safety stock standard'.

[0124] For example, the device (200) can be configured to automatically request an order when the current inventory of materials is less than or equal to a preset safety stock quantity (e.g., less than 1.5 times the average usage over the last 30 days).

[0125] At this time, the pre-set safety stock standard may be set differently depending on the importance of the material, the risk of delivery delay, and past usage variability, and this is merely an example and may vary depending on the embodiment.

[0126] For example, the device (200) can be configured to pre-order replacement parts such as pump filters, valves, and O-rings, taking into account the delivery date until the predicted maintenance time, when the abnormal sign score of the cooling pump is 0.8 or higher.

[0127] At this time, the pre-set period may be set within a range of 3 to 14 days based on material delivery dates, equipment maintenance cycles, material inventory turnover rates, etc., and this is merely an example and may vary depending on the embodiment.

[0128] Through this, the device (200) can prevent parts shortages or unnecessary excess inventory and secure a material supply system aligned with the maintenance schedule.

[0129] In step S205, the device (200) can adjust grouped complex-specific operational plans, including security schedules, cleaning schedules, and common area management schedules, based on facility status prediction information according to preset criteria.

[0130] That is, the device (200) can automatically reassign security, cleaning, and common area management schedules among the operation plans of each complex based on facility status prediction information.

[0131] The device (200) can prioritize the placement of inspection personnel in areas where there is a high probability of equipment malfunction and optimize work schedules by taking into account the availability of personnel for each complex.

[0132] Additionally, the device (200) can analyze the work schedule, movement path, and work difficulty data of each employee and apply a schedule optimization algorithm (e.g., genetic algorithm, constraint-satisfying scheduling algorithm).

[0133] Through this, the device (200) can increase the frequency of work placement in high-risk areas while minimizing the total travel distance and overlapping work time of the personnel.

[0134] At this time, the pre-set criteria may be set so that the work efficiency index is 0.8 or higher, or so that the total daily movement path is limited to 3 km or less, and this is merely an example and may be set differently depending on the embodiment.

[0135] For example, the device (200) can adjust the schedule by increasing the frequency of patrols by cleaning and security personnel in areas near electrical rooms with high abnormal sign scores from 2 to 4 times a day, and reducing the frequency of patrols in areas with relatively low risk.

[0136] At this time, the pre-set criteria may be established by considering the working hours of personnel, movement paths, work efficiency index, priority of hazardous areas, etc., and this is merely an example and may be set differently depending on the embodiment.

[0137] Through this, the device (200) can maximize the work efficiency of management personnel and increase the level of response to dangerous areas, thereby improving the stability of the complex's operation.

[0138] In this regard, a detailed explanation will be provided in Fig. 5.

[0139] In step S206, the device (200) can automatically generate grouped complex accounting management information including maintenance costs, labor costs, material costs, and reserve costs based on facility status prediction information and grouped complex operation plans.

[0140] That is, the device (200) can automatically calculate accounting management information including maintenance costs, labor costs, material costs, and reserve costs by combining facility status prediction information and operation plans for each complex.

[0141] The device (200) can predict the cost for each item and determine whether it is exceeded by comparing it with the budget limit.

[0142] In addition, the device (200) can calculate the predicted value of each item by referring to the history of budget execution by complex, material unit price fluctuation trends, and labor unit price history data stored in the accounting management database.

[0143] For example, when predicting maintenance costs, the device (200) can estimate the required manpower input time by multiplying the average work time per facility by the work difficulty coefficient and calculate labor costs by applying the labor cost unit price.

[0144] Material costs can be calculated by reflecting the average of recent unit prices by item in the material management database and delivery adjustment costs, and reserve costs can be calculated by applying a preset ratio (e.g., 3~10%) to the total sum of each item.

[0145] The device (200) can determine whether the total amount of the calculated accounting item is exceeded by comparing it with a preset budget limit, and if it is exceeded, it can automatically generate 'budget exceedance alarm data'.

[0146] At this time, the pre-set budget limit may be set by considering the total number of households per complex, the average of the previous year's management fees, the monthly fluctuation rate, etc., and this is merely an example and may be set differently depending on the embodiment.

[0147] For example, if the maintenance work time for the device (200) is predicted to be 6 hours, the maintenance cost can be calculated by adding the labor cost unit price (e.g., 20,000 won per hour) and the material cost (e.g., 80,000 won), and the total cost can be calculated by applying a reserve ratio (e.g., 5%) to it.

[0148] At this time, the pre-set reserve ratio may be set within the range of 3 to 10% considering management regulations, the size of the complex, and seasonal maintenance frequency, and this is merely an example and may be set differently depending on the embodiment.

[0149] Through this, the device (200) can ensure transparency in budget execution and prevent overspending of management costs.

[0150] For a detailed explanation regarding this, refer to Fig. 6.

[0151] In step S207, the device (200) can integrate, store, and manage grouped complex-specific management data including facility status prediction, material ordering, operation plan coordination, and accounting management.

[0152] That is, the device (200) can manage management data by integrating information generated from equipment status prediction, material ordering, operation planning, accounting management, etc.

[0153] The device (200) can store the data produced at each stage in an integrated management database and provide it in a visualized form to the terminal (100) of the apartment building manager.

[0154] In addition, the device (200) can store equipment status data, material order history, operation plan schedule, and accounting item data in an interlocking table of an integrated management database.

[0155] The integrated management database can be indexed with complex-specific identifiers (IDs), unique equipment codes, material item codes, accounting item codes, etc., and configured to allow querying and updating of management data at the complex and group levels.

[0156] The device (200) maintains the latest state by setting the data update cycle to a preset standard (e.g., real-time, 5 minutes, 1 hour), and can control data viewing, modification, and approval according to user authority when the terminal (100) of the apartment building manager accesses it.

[0157] At this time, the preset update cycle and access rights may be set merely by considering the scale, network load, data importance, etc., and this is merely an example and may be set differently depending on the embodiment.

[0158] For example, the device (200) displays the facility risk level, material order status, and budget execution rate by complex in the form of graphs and tables on a single dashboard screen, and the manager can view the information in real time through the apartment building manager's terminal (100).

[0159] Through this, the device (200) can comprehensively manage the operational status of each complex and strengthen the decision support system for the manager.

[0160] FIG. 3 is a flowchart illustrating the process of grouping a multi-unit housing complex according to one embodiment.

[0161] Referring to FIG. 3, first, in step S301, the device (200) can generate a first group for maintenance efficiency based on the degree of equipment deterioration, equipment type, and number of households per apartment complex.

[0162] That is, the device (200) analyzes the facility management history, design drawing information, and household composition data stored in the multi-unit housing database to calculate an index that combines the degree of facility obsolescence (years of use / standard lifespan per facility), the type of facility (major facility categories such as boilers, elevators, fire pumps, etc.), and the number of households (complex size indicator), and can classify complexes requiring maintenance efficiency into a first group according to pre-set criteria.

[0163] At this time, the device (200) normalizes each item to enable comparison between items and can apply weights according to the importance of each item.

[0164] For example, the device (200) can designate a complex as the first group that has an equipment obsolescence index of 0.7 or higher, has a proportion of high-risk equipment such as elevators and boilers of 40% or more, and has 300 or more households.

[0165] At this time, the pre-set criteria may be established by combining standard life data, equipment risk classification table, complex size ranges (e.g., small <300, medium 300~700, large >700), and periodic inspection cycles according to management regulations, and this is merely an example and may be set differently depending on the embodiment.

[0166] Additionally, the device (200) may be configured to periodically update the relevant standards according to changes in management policies or equipment replacement history.

[0167] Through this, the device (200) can select complexes with a high proportion of old, large-scale, and high-risk facilities and prioritize the deployment of resources for regular inspection and preventive maintenance, thereby maximizing maintenance efficiency.

[0168] In step S302, the device (200) can generate a second group for reducing management costs based on energy usage and maintenance cost patterns for each apartment complex.

[0169] That is, the device (200) can analyze the energy usage (kWh, in units of heat) and maintenance cost details (monthly or quarterly expenditure, maintenance cost per unit area) by household and by complex stored in the multi-unit housing database, and classify complexes with low energy-to-cost efficiency into a second group.

[0170] The device (200) can normalize the collected data into a cost per unit household, calculate a correlation coefficient between energy usage and maintenance costs, and automatically identify low-efficiency complexes according to preset criteria.

[0171] For example, the device (200) may designate a complex as a second group if the energy usage per household over the past 12 months is in the top 30% and the maintenance cost per household is in the top 30% or higher, or if the maintenance cost per energy indicator (maintenance cost / kWh) is 1.2 times or more than the average of the complex.

[0172] At this time, the pre-set criteria may be set based on the analysis period (e.g., 6 to 24 months), upper and lower quantiles (e.g., 20%, 30%, 40%), the weighted average method of the indicator, standard guidelines for management costs, or energy efficiency statistics from the National Energy Agency, etc., but this is merely an example and may be set differently depending on the embodiment.

[0173] Additionally, the device (200) can generate improvement feedback data that can adjust equipment operation conditions, inspection cycles, and contract unit price information for the second group complex and provide it to the manager's terminal (100).

[0174] Through this, the device (200) can identify complexes with energy-cost inefficiency sections and systematically carry out strategic measures to reduce management costs, such as facility adjustment, operation improvement, and contract change.

[0175] In step S303, the device (200) can generate a third group for service operation efficiency based on the age group, income level, and lifestyle patterns of residents in each apartment complex.

[0176] That is, the device (200) can analyze demographic data of residents (e.g., age group, household composition type, average income per household, length of residence, etc.) and lifestyle pattern data (e.g., frequency of use of common facilities, time of day for night lighting use, type of complaint received, etc.) stored in a multi-unit housing database to derive service demand characteristics for each complex and form a third group accordingly.

[0177] The device (200) can automatically classify the type of complex for service operation efficiency by analyzing the correlation between the residents' living time zones by age group, the frequency of use of common facilities, and the types of service complaints.

[0178] For example, the device (200) can classify a complex with a high proportion of elderly people (40% or more) and high night lighting usage as a ‘night safety service enhanced complex,’ and a complex with a high proportion of dual-income households in their 30s and 40s (60% or more) and low daytime common facility usage rate as a ‘non-working hours intensive management type complex.’

[0179] At this time, the pre-set criteria may include age groups (e.g., young adults aged 20–39, middle-aged adults aged 40–59, elderly adults aged 60 or older), household types (single-person households, family households, etc.), income quintiles (top 30%, middle 40%, bottom 30%), and usage rates of public facilities (e.g., a difference of 10% or more in daytime and nighttime usage rates), but this is merely an example and may be set differently depending on the embodiment.

[0180] In addition, the device (200) can set service provision priorities for each third group and automatically generate a customized service plan in conjunction with cleaning, security, and facility management schedules.

[0181] Through this, the device (200) can establish a customized operation system for the complex that reflects the characteristics and lifestyle patterns of the residents, thereby improving management efficiency and resident satisfaction at the same time.

[0182] In step S304, the device (200) can set data processing weights corresponding to the management purpose for each group based on the first group, the second group, and the third group.

[0183] That is, the device (200) sets data processing weights to be referenced in the analysis and decision-making process according to the management purpose for each group (1st group: maintenance efficiency, 2nd group: cost reduction, 3rd group: service operation efficiency), and can reflect the corresponding weights in subsequent steps (equipment status prediction, material ordering, operation planning, accounting calculation).

[0184] Here, data processing weights are coefficients ranging from 0 to 1, and refer to values ​​multiplied by the loss function, regression weight, or priority formula of an artificial intelligence model to adjust the influence of each analysis item.

[0185] For example, the device (200) may assign a weight of 0.6 for equipment condition prediction accuracy, a weight of 0.2 for shortening the inspection cycle, and a weight of 0.2 for material procurement stability to the first group, a weight of 0.5 for reducing maintenance costs, a weight of 0.3 for improving energy efficiency, and a weight of 0.2 for optimizing material unit prices to the second group, and set a weight of 0.4 for responding to peak hours, a weight of 0.3 for mitigating complaints, and a weight of 0.3 for operating efficiency of public facilities to the third group.

[0186] These weights can be determined by comprehensively considering the size of each complex, recent operational performance indicators, management goals (e.g., 10% cost reduction, 20% reduction in complaints), and the results of the AI ​​model's learning feedback.

[0187] In addition, the device (200) can adjust the centrality of the group-specific analysis results based on the set weights to produce prediction and operation results optimized for each of the maintenance-centered complex, cost-reduction-centered complex, and service-efficiency-centered complex.

[0188] Through this, the device (200) can improve the efficiency and precision of overall complex management by implementing a differentiated decision-making system that reflects the characteristics and management direction of each group.

[0189] FIG. 4 is a flowchart illustrating the process of generating equipment status prediction information according to one embodiment.

[0190] Referring to FIG. 4, first, in step S401, the device (200) can extract a pattern of change of current value, temperature value, pressure value and vibration value from equipment status data.

[0191] That is, the device (200) can quantitatively identify the fluctuation trends and abnormal changes of each physical variable by analyzing the hourly changes of current, temperature, pressure, and vibration collected in real time from IoT sensors installed in each facility within the apartment complex.

[0192] The device (200) can divide the collected data into intervals at regular time intervals, calculate statistical characteristics such as average value, maximum value, minimum value, standard deviation, rate of change, and periodicity in each interval, and calculate a change pattern through this.

[0193] For example, the device (200) can calculate the rate of change of the current value (rate of change relative to the previous interval), the rate of temperature rise (amount of temperature increase per unit time), the coefficient of variation of the pressure value (value obtained by dividing the standard deviation by the mean value), and the energy distribution of the vibration signal by frequency (e.g., 10~30 Hz, 30~80 Hz) by setting time intervals of 1 minute, 5 minutes, and 60 minutes.

[0194] Additionally, the device (200) may perform refinement procedures such as missing value correction, removal of abnormal data, and truncation of excess values ​​(e.g., removal of values ​​exceeding the mean ±3 times standard deviation range) to remove outliers or noise that may be included in the sensor data.

[0195] Here, the length of the time interval, the frequency analysis range, and the outlier handling criteria may be set differently depending on the type of equipment, the data collection cycle, and the operating conditions; this is merely an example and may be set differently depending on the embodiment.

[0196] Through this, the device (200) can secure change characteristics over time intervals reflecting the operating status of each facility and use them as input data for determining abnormal signs and predicting maintenance in the next stage.

[0197] In step S402, the device (200) inputs the change pattern of the extracted equipment status data into an artificial intelligence model and calculates an abnormal sign score according to a preset standard by comparing it with a past normal state pattern.

[0198] That is, the device (200) can use the change pattern as input to an artificial intelligence model (e.g., a supervised learning-based SVM classifier or an autoencoder for outlier detection) and calculate an anomaly sign score by calculating the deviation from the defined normal state pattern from the training data collected in the normal operation section for each facility.

[0199] The device (200) comprehensively evaluates the similarity between patterns (e.g., cosine similarity, Euclidean distance), temporal trend difference (e.g., moving average deviation), frequency spectrum difference (e.g., power density difference), etc. to generate a continuous abnormality score in the range of 0 to 1, and if it exceeds a certain range, it can classify it as a high-risk range.

[0200] For example, the device (200) can recalculate the final score by adding +0.1 to the abnormal sign score when two or more of the following occur simultaneously: vibration spectrum deviation (cosine distance greater than 0.25), temperature rise rate (greater than 2℃ / h), current spike frequency (greater than 3 times within 10 minutes), and pressure fluctuation coefficient (greater than 0.15).

[0201] At this time, the threshold, weight, and boundary values ​​may be set based on statistical reference values ​​(e.g., mean ±2σ, 95% confidence interval) calculated from the training dataset of the normal operating range used to train the AI ​​model or the allowable specification standards of the equipment manufacturer, and this is merely an example and may be set differently depending on the embodiment.

[0202] Additionally, the device (200) can record the calculated abnormal sign scores by facility and time and provide them as input variables to predict when maintenance is needed in a subsequent step.

[0203] Through this, the device (200) can quantitatively derive an early warning signal of equipment failure by comprehensively analyzing pattern fluctuations of multiple indicators beyond the level of anomaly detection of a single sensor.

[0204] In step S403, the device (200) can calculate the probability of an anomaly occurring according to preset criteria by combining the calculated anomaly sign score with the equipment-specific aging index, usage frequency, and recent inspection cycle.

[0205] That is, the device (200) can calculate the probability of an abnormal occurrence (probability value or grade) by combining an abnormal sign score with an auxiliary indicator consisting of an equipment aging index (a value of 0 to 1 converted from years of service / standard lifespan, etc.), usage frequency (operation time / number of starts, etc.), and recent inspection cycle (number of days elapsed since the last inspection / recommended inspection cycle).

[0206] For example, the device (200) can calculate P = 0.4 × abnormal signs + 0.3 × aging + 0.2 × usage frequency + 0.1 × recent inspection cycle for equipment with an abnormal sign score of 0.6, an aging index of 0.8, a usage frequency of 0.7, and a recent inspection cycle of 0.9 by applying the weighted sum P = 0.4 × abnormal signs + 0.3 × aging + 0.2 × usage frequency + 0.1 × recent inspection cycle, and can classify 0.7 or higher as “high probability of abnormal occurrence”.

[0207] At this time, the weights and boundary values ​​may be set based on the recommended inspection cycle provided by the manufacturer, the average failure cycle per piece of equipment, or the results of the analysis of past equipment maintenance history data; however, this is merely an example and may be set differently depending on the embodiment.

[0208] Through this, the device (200) can provide a realistic possibility assessment that reflects both the current abnormal signal and the structural and operational risk factors of the equipment.

[0209] In step S404, the device (200) can predict the timing of maintenance requirements and the scope of maintenance work by applying a preset equipment importance weight to the calculated probability of anomaly occurrence.

[0210] That is, the device (200) can calculate the timing of maintenance required (e.g., immediate / short-term / medium-term) and the scope of maintenance work (inspection, parts replacement, overhaul, etc.) by reflecting the equipment importance weight (a value of 0 to 1 calculated based on safety impact, service interruption impact, whether subject to statutory inspection, etc.) to the possibility of an abnormality occurring.

[0211] For example, the device (200) can be set to calculate a risk score R = probability × importance = 0.585 for a fire pump (equipment importance weighting 0.9) with a probability of 0.65 of an abnormality, and to determine the scope of work as “short-term (e.g., within 7 days) inspection and replacement of packings” if R is 0.55 or higher, and “immediate inspection and replacement of core parts” if R is 0.75 or higher.

[0212] At this time, the weight formula, critical region, and task classification are merely examples and may be set differently depending on the embodiment.

[0213] For example, 'immediate inspection' may mean within 3 days, 'short-term inspection' within 7 days, and 'medium-term inspection' within 30 days, and these may be set differently depending only on the scale and importance of the equipment.

[0214] Through this, the device (200) can derive a maintenance plan that differentiates the timing and scope to prioritize the handling of critical facilities for safety and continuity, even if the possibilities are the same.

[0215] In step S405, the device (200) generates equipment condition prediction information including the likelihood of anomalies, the time when maintenance is required, and the scope of maintenance work, and can reflect the generated equipment condition prediction information in the grouped management data by complex.

[0216] That is, the device (200) can generate equipment status prediction information (a record consisting of the probability of anomaly occurring, the time when maintenance is required, and the scope of maintenance work) for each piece of equipment, and store and link this to grouped complex-specific management data (a data set for operation, material, and accounting linkage integrated at the complex, group, and equipment unit level).

[0217] For example, the device (200) can create a record of the facility ID, the probability of an abnormality occurring (0.71), the time when maintenance is required (short-term: 7 days), and the scope of maintenance work (replacement of spare parts + detailed inspection), and automatically connect to the material management database and accounting management information of the complex to simultaneously update material orders (e.g., 2 filters, 1 set of packing) and cost predictions (e.g., labor costs, material costs, reserves).

[0218] At this time, the linkage structure and field configuration are merely examples and may be set differently depending on the embodiment.

[0219] The device (200) can be configured to store the generated equipment status prediction information in a complex-specific integrated management database or a central management server, and to enable real-time reflection by performing data synchronization with the material management and accounting management systems at one-minute intervals.

[0220] Through this, the device (200) can implement a data flow in which the prediction results are immediately propagated to material ordering, operation planning, and accounting calculations, thereby improving the accuracy and agility of integrated management by complex.

[0221] FIG. 5 is a flowchart illustrating the process of coordinating operation plans for each grouped complex according to one embodiment.

[0222] Referring to FIG. 5, first, in step S501, the device (200) can calculate the risk level of each piece of equipment based on the predicted probability of an abnormality occurring in the equipment and the time when maintenance is required.

[0223] That is, the device (200) can receive the possibility of anomalies occurring for each facility and the time when maintenance is required as input, and calculate a risk level that reflects the impact on the safety and service continuity of the facility.

[0224] Here, the risk level is a comprehensive judgment indicator calculated by the device (200) by combining the likelihood of failure of the equipment, the urgency of maintenance, and the impact on service, and can be defined as a normalized value in the range of 0 to 1.

[0225] For example, the device (200) can be calculated as risk R = 0.5 × (probability of anomaly) + 0.3 × (urgency index) + 0.2 × (impact index), and the urgency index can be defined based on the maintenance policy of the complex, the inspection cycle recommended by the equipment manufacturer, or the safety management manual.

[0226] For example, it can be set to 'immediately: 1.0 / within 7 days: 0.8 / within 30 days: 0.5 / more than 30 days: 0.2', and the impact index can be defined as 0.9 for fire safety and elevators, etc., and 0.5 for lighting and ventilation facilities, etc., considering the safety impact of the facility, the convenience of residents' living, and whether it is subject to statutory inspection.

[0227] At this time, the formula, coefficient, and interval are merely examples and may be set differently depending on the embodiment.

[0228] Through this, the device (200) can obtain quantitative indicators that reflect the magnitude and urgency of the risk for each facility, and use them as a basis for setting priorities for security, cleaning, and inspection tasks and adjusting work schedules in subsequent stages.

[0229] In step S502, the device (200) can set priorities for security, cleaning, and common area management tasks based on the calculated risk level.

[0230] That is, the device (200) can classify the risk level of each zone within the complex according to the risk level of the facility and set work priorities so that more management resources can be allocated to zones with high risk.

[0231] Here, priority can be defined as a criterion determined by comprehensively considering the relative risk level, the need for management, safety impact, and the frequency of recent anomalies.

[0232] For example, the device (200) can classify the top 20% of the risk level by zone into 'Priority Zone A', the 20-50% into 'Priority Zone B', and the 50% or less into 'Priority Zone C', and can set the priority zone A to perform patrols at least 4 times a day, cleaning at least 2 times, and inspection at least 1 time, and the priority zone B to perform patrols 2 times and cleaning 1 time, and the priority zone C to perform patrols 1 time as a basic standard.

[0233] At this time, the ratio of each section division or the frequency of management may be adjusted considering the size of the complex, the number of resident households, facility density, and statutory inspection standards; however, this is merely an example and may be set differently depending on the embodiment.

[0234] In addition, if multiple zones with the same level of risk exist, priority can be determined by using additional criteria such as recent complaint history, accident frequency, and nighttime safety light inspection status.

[0235] Through this, the device (200) can simultaneously improve the operational efficiency of management personnel and the safety of the complex by efficiently distributing more management personnel and inspection resources to high-risk areas within the complex.

[0236] In step S503, the device (200) can automatically generate a work schedule by adjusting the placement of security personnel, cleaning personnel, and inspection personnel based on a set priority and according to a preset standard, taking into account the estimated time required for each task, the number of available personnel, and movement paths.

[0237] That is, the device (200) assigns the necessary personnel and work frequency starting from high-priority zones, and can automatically calculate a work schedule by comprehensively calculating the estimated time required for the work and the available working hours of the personnel.

[0238] The criteria set in advance here serve as judgment criteria for generating work schedules and may include items such as the maximum daily working hours per person, the allowable continuous working hours, the night shift ratio, movement restrictions in restricted areas, the minimum waiting interval between inspection areas, the risk level of each area, and the importance of the work. These criteria may be set by considering the size of the complex, management regulations, personnel deployment policies, seasonal operational characteristics, etc., and this is merely an example and may be set differently depending on the embodiment.

[0239] The device (200) can apply the formula “T = basic work time + travel time + waiting or setting time” to calculate the estimated time required for the work.

[0240] The basic work time may vary depending on the difficulty of the work in the area, the number of work items, and whether equipment is used; for example, in areas with more inspection items than simple patrols, the basic work time may be calculated to be longer.

[0241] Travel time can be calculated by considering the physical distance between zones, passable routes, elevator waiting times, and whether access to controlled zones is restricted.

[0242] Waiting or setup time may include preparation time before starting work, equipment inspection time, and cleanup time after finishing work.

[0243] The device (200) can distribute the estimated time calculated in this way to the overall work schedule by reflecting the available working hours, shift system, and rest time of the personnel. In addition, the device (200) can analyze the shortest movement path between zones to optimize movement paths and automatically search for a path that avoids closed zones or congested areas to readjust personnel placement.

[0244] In this process, the device (200) can secure realistic feasibility by simultaneously reviewing constraints such as whether there is a conflict in the schedule, the possible continuous working hours, and the ratio of night work hours.

[0245] Through this, the device (200) can automatically derive a feasible work schedule that goes beyond simply assigning tasks and reflects both actual physical travel time and manpower availability conditions, thereby minimizing unnecessary travel or duplicate tasks and improving operational efficiency and response speed.

[0246] In step S504, the device (200) can calculate an efficiency index for each task according to preset criteria to evaluate the management efficiency of the generated work schedule.

[0247] That is, the device (200) can calculate an efficiency index for each task and use the result as a schedule verification indicator to quantitatively evaluate how reasonably the automatically generated work schedule is structured in terms of time, manpower, and risk.

[0248] Here, the efficiency index by task is an evaluation indicator that quantifies the quality of work schedules, and can be calculated as a value between 0 and 1 by combining the ratio of on-site work time, the ratio of travel time, the suitability of input time relative to risk, and the workforce utilization rate.

[0249] An efficiency index closer to 1 indicates that actual working time is concentrated on on-site work and unnecessary movement or waiting time is minimized, whereas a value closer to 0 indicates high inefficiency of the schedule.

[0250] For example, the device (200) can calculate an efficiency index according to the formula 'E = 0.4 × (field work time ÷ total working time) + 0.2 × (1 travel time ÷ total working time) + 0.3 × (risk-weighted input time suitability) + 0.1 × (manpower utilization rate)'.

[0251] Here, the first term (on-site work time ÷ total working time) represents the proportion of work actually performed on-site out of the total work, and a higher value indicates a higher concentration of the schedule.

[0252] The second term (1 travel time ÷ total working hours) reflects the proportion of travel time in total working hours and serves to correct for inefficiencies caused by travel.

[0253] The third term, “risk-weighted input time suitability,” is an indicator of whether sufficient time has been allocated to high-risk zones. For example, it can be evaluated as 1.0 if 40% or more of the total input time is allocated to high-risk zones, and 0.6 if less.

[0254] The last term, “personnel utilization rate,” indicates how efficiently each employee is allocated relative to their actual working hours and can be calculated by taking into account break times, standby times, etc.

[0255] At this time, the weights of each item (0.4, 0.2, 0.3, 0.1) may be set differently depending on the purpose of the work schedule and the operational characteristics of the complex, and this is merely an example and may be set differently depending on the embodiment. In addition, the standard value of the efficiency index (e.g., 0.75 or higher is appropriate, and less than 0.75 is subject to readjustment) may be set as the minimum quality standard for management efficiency.

[0256] Through this, the device (200) can not only generate a simple schedule but also quantitatively verify whether the work schedule is actually efficient and structured around risk response.

[0257] As a result, the device (200) can simultaneously improve the quality of the work schedule and the efficiency of on-site response by objectively verifying whether the proportion of actual work relative to time and the concentrated distribution of the danger zone have been secured.

[0258] In step S505, the device (200) can regenerate the work schedule if the efficiency index is less than a preset threshold value.

[0259] That is, if the calculated efficiency index is evaluated to be below the management standard, the device (200) can automatically re-search for a new work schedule after adjusting one or more of the priority weights, movement path constraints, and personnel placement rules to improve the quality of the schedule.

[0260] Here, the preset threshold value refers to the minimum quality standard of the work schedule and can be defined as a critical line for determining whether the schedule satisfies the balance between operational efficiency and risk response.

[0261] For example, the standard value can be set to an appropriate level when the efficiency index (E) is 0.75 or higher, but the standard value can be raised to 0.8 in cases where the scale is large or the complexity of the equipment is high, and conversely, it can be relaxed to 0.7 in cases of short-term inspections or schedules centered on low-risk sections.

[0262] These standard values ​​may be set differently depending on the operational characteristics of the complex, seasonal factors (e.g., winter / summer equipment load), or special schedules (e.g., regular inspection weeks, external audit schedules).

[0263] The device (200) can perform a readjustment step to improve efficiency when the reference value is not met.

[0264] For example, if the initial efficiency index is calculated as 0.71, the device (200) can increase the risk response ratio by increasing the minimum input time ratio of the high-risk zone by +10%, reduce unnecessary movement by shortening the maximum allowable time of the movement path by 15%, and improve the dispersion of personnel by relaxing the upper limit of night shift concentration.

[0265] Such adjustments can be performed repeatedly by gradually relaxing or tightening the constraints of the schedule generation algorithm, and the adjustment range and order of each item can be determined according to a predefined adjustment rule table.

[0266] At this time, the reference value, adjustment range, priority adjustment item, and number of repetitions are merely examples and may be set differently depending on the actual embodiment.

[0267] Through this, the device (200) can implement a self-correcting schedule management function that automatically improves quality until the efficiency index reaches the target standard, going beyond a simple schedule generation level.

[0268] As a result, the device (200) can optimize itself so that the work schedule actually ensures operational efficiency and maintains a risk-response-centered resource allocation structure.

[0269] In step S506, the device (200) can adjust work schedules in which the efficiency index is greater than or equal to a preset threshold value by reflecting them in the grouped operation plan.

[0270] That is, the device (200) classifies work schedules in which the efficiency index is evaluated to be above a standard value as “verification completed schedules” and can reflect them as confirmed schedules of grouped complex-specific operation plans.

[0271] At this time, the confirmed and reflected schedule allows for the systematic management of schedule history by recording the schedule's version information, application period, application area, and related personnel information together.

[0272] The device (200) can transmit a confirmed work schedule to the terminal (100) of the apartment building manager and provide a function that allows the manager to check the schedule or request modifications if necessary through the terminal.

[0273] For example, the device (200) can create a schedule file including metadata such as “week of November 3, 2025 / Group 1 complex / electrical room and machine room center work schedule” and transmit it to the terminal (100) of the apartment building manager.

[0274] The terminal (100) of the apartment building manager visually displays schedule data received from the device (200) and can provide a task list, person in charge, time zone, travel route, etc. in the form of a table, map, or Gantt chart.

[0275] Additionally, the manager can check the execution status of the schedule in real time through the terminal or input a request for modification according to the on-site conditions, and such a request can be transmitted back to the device (200) and reflected in the schedule adjustment process.

[0276] At this time, the device (200) stores metadata of each schedule in a management database and records together the “schedule version (Version ID),” “applicable complex group ID,” “confirmed date,” and “application period (e.g., 2025-11-03~2025-11-09)” so that it can be used for future schedule change history or performance analysis.

[0277] At this time, the display format, metadata configuration, visualization method, etc., may be set differently depending on the embodiment.

[0278] Through this, the device (200) can maintain consistency in field operations and minimize unnecessary schedule confusion by reflecting only schedules that meet efficiency standards in the grouped complex operation plan.

[0279] In addition, through the terminal (100) of the apartment building manager, the manager can clearly check the confirmed schedule and intuitively understand the execution status, thereby ensuring both efficiency and transparency in operation management.

[0280] FIG. 6 is a flowchart illustrating the process of automatically generating accounting management information for grouped complexes according to one embodiment.

[0281] Referring to FIG. 6, first, in step S601, the device (200) can determine the time required for maintenance and the maintenance work range for each piece of equipment based on equipment condition prediction information, and determine the time required for the work by looking up a preset work time value corresponding to the work range.

[0282] That is, the device (200) can identify specific maintenance steps required for each piece of equipment and the timing of their execution by analyzing data on the probability of anomaly occurring for each piece of equipment, the timing of maintenance required, and the scope of maintenance work included in the equipment condition prediction information.

[0283] The device (200) can look up preset work time values ​​corresponding to the conditions in a standard work time table stored internally, taking into account the difficulty and scope of maintenance work for each facility.

[0284] Here, the preset work time value refers to a predefined average execution time calculated by combining factors such as equipment type, work scope, environmental conditions (e.g., indoor / outdoor, accessibility, etc.), and difficulty level; this can be set based on standard maintenance manuals for each piece of equipment, historical maintenance data, or management regulations.

[0285] For example, if the device (200) predicts that the maintenance time required for the “elevator (equipment ID: E-102)” is ‘short-term’ and the maintenance work scope is ‘part replacement’, it can look up a preset standard work time value of 2.8 hours and determine the actual work time required as 2.9 hours by reflecting field correction factors (e.g., access time due to high-rise building +0.3 hours, parallel work within the same area 0.2 hours).

[0286] At this time, the preset work time values ​​and correction factors are merely examples and may be set differently depending on the embodiment.

[0287] Through this, the device (200) can calculate the time required for work according to the timing of maintenance needs and the scope of work in a systematic and consistent manner, and can increase the accuracy of the subsequent labor cost, maintenance cost, and work schedule calculation steps.

[0288] In step S602, the device (200) can check the work schedule for each personnel based only on the operation plan and calculate labor costs based on the work time and preset labor cost unit price.

[0289] That is, the device (200) can retrieve the work schedule of the personnel assigned to each task in the finalized operation plan and calculate the actual input time including the work start time, end time, work time (day, night, holidays, etc.), and rest and waiting periods.

[0290] The device (200) can refer to a preset labor cost table defined according to each person's job group (e.g., security, cleaning, inspection, maintenance), skill level (e.g., beginner, intermediate, advanced), and working time (e.g., day, night).

[0291] The pre-set labor cost unit price here refers merely to an hourly or daily rate established based on management regulations, labor standards, or past payroll data, and may include a differentiated rate system by time of day or job category.

[0292] For example, in the case of the same work where “inspection personnel A (daytime unit price 22,000 won / hour, intermediate skill level)” is deployed for 1.2 hours and “maintenance personnel B (nighttime unit price 28,000 won / hour, advanced skill level)” is deployed for 1.7 hours, the device (200) can calculate the labor cost for A as 26,400 won (22,000 × 1.2) and the labor cost for B as 52,360 won by applying a nighttime additional rate of 10% to 47,600 won (28,000 × 1.7), and calculate the labor cost per work as 78,760 won by summing these.

[0293] At this time, the pre-set unit price, addition rate, rounding rules, etc. are merely examples and may be set differently depending on the embodiment.

[0294] In addition, the device (200) can automatically adjust overlapping times or distribute labor costs by reflecting the ratio of parallel work when there are multiple tasks within the same work schedule.

[0295] Through this, the device (200) can automatically calculate labor costs by combining work schedules and labor cost unit prices, thereby standardizing the calculation process of labor input costs and increasing the accuracy of the accounting linkage stage.

[0296] In step S603, the device (200) can calculate maintenance costs based on a pre-set unit price per task, based on the predicted maintenance work range.

[0297] That is, the device (200) can calculate maintenance costs by referring to a preset unit price table for each task, in which the task type, difficulty, and equipment grade are mapped according to the maintenance task scope (e.g., inspection, light maintenance, parts replacement, overhaul) specified in the equipment condition prediction information.

[0298] Here, the pre-set unit price per task refers to a standard value defined by analyzing maintenance history by complex, past outsourcing contract unit prices, and data on manpower and time required for each standard process.

[0299] This unit price may be set differently depending on the type of equipment (e.g., boiler, fire pump, elevator, etc.), the difficulty of the work (e.g., general, high difficulty, requiring special equipment), and the equipment grade (e.g., small, medium, large).

[0300] In addition, the device (200) can calculate maintenance costs subdivided by each item by defining a rule on whether to include labor costs, equipment usage fees, travel expenses, and incidental expenses (e.g., tool and consumable costs).

[0301] For example, the device (200) can apply unit price rules such as “Boiler / Light Maintenance: 85,000 won / case (Large Equipment +10%)”, “Fire Pump / Overhaul: 600,000 won / case (Use of Special Equipment +5%)”, “Elevator / Parts Replacement: 120,000 won / case (High Difficulty +15%)”.

[0302] In addition, if continuous maintenance work is scheduled for the same facility, a 10% discount rate may be applied to reflect the labor cost savings resulting from bundled work, and such unit price adjustment rules are merely examples and may be set differently depending on the embodiment.

[0303] Through this, the device (200) can calculate standardized maintenance costs corresponding to the scope of maintenance work and manage components such as labor costs, materials, and equipment costs according to consistent rules, thereby increasing the accuracy and ease of verification of accounting information calculation.

[0304] In step S604, the device (200) can calculate the material cost by checking the preset material unit price based on the order information of the generated material.

[0305] That is, the device (200) can automatically calculate material costs by referring to the item name, specifications, quantity, delivery date, and supplier information included in the material order information, and by querying a preset material unit price table defined by item and specifications.

[0306] The pre-set material unit price here refers to the standard unit price stored in the material management database, meaning a reference value calculated by combining past contract unit prices, public notices from the Public Procurement Service or standard unit prices, unit prices from recent delivery records, and bundled transaction unit prices.

[0307] In addition, the device (200) can reflect whether to include incidental costs such as transportation costs, taxes, packaging costs, and warehouse storage fees in addition to the material unit price, according to a pre-set calculation rule.

[0308] For example, the device (200) can look up standard unit prices for materials such as “filter cartridge (standard A) 2EA × 25,000 won / EA, pump packing set 1 set × 18,000 won / set (2 sets reflected when applying the minimum order 2 sets rule), lubricating oil 8 liters × 8,000 won / liter (5% bundle unit price not applied when purchasing 10 liters or more)”.

[0309] The device (200) can calculate the final material cost by applying an urgent ordering surcharge (e.g., +7%) when a request for a shortened delivery time is included, and conversely, by reflecting a bundle discount rate (e.g., 5%) when bulk purchase conditions are met.

[0310] In this case, the markup rate, discount rate, and minimum order conditions are merely examples, and the actual application criteria may vary depending on the material type, supply contract policy, order quantity range, etc.

[0311] Through this, the device (200) can match the material order information with the preset material unit price and automatically calculate the material cost reflecting the delivery date, transaction conditions, and incidental costs, thereby preventing under- or over-calculation errors and increasing the accuracy of budget execution.

[0312] In step S605, the device (200) can calculate a reserve cost by applying the reserve ratio to the sum of maintenance costs, labor costs, and material costs based on a preset reserve ratio.

[0313] That is, the device (200) can calculate a reserve by applying a preset reserve ratio defined by management regulations or accounting policies per complex to the sum of costs (maintenance costs + labor costs + material costs) of each work unit.

[0314] The pre-set reserve ratio here refers to a ratio determined by comprehensively considering the size of the complex, facility complexity, seasonal maintenance risk, accessibility, etc. For example, the default value is 5%, and for high-risk complexes (top 20% in facility density and obsolescence), +2% can be added, and for winter work when the outside temperature is 5℃ or lower, +1% can be added.

[0315] The device (200) manages these reserve ratio setting rules in the form of a standardized reference table, so that it can correct them to ensure that consistent reserve ratio calculations are always made for work under the same conditions.

[0316] For example, the device (200) can calculate a reserve of 59,740 won by applying a total reserve ratio of 8% (basic 5% + winter 2% + high-risk equipment 1%) to the sum of maintenance costs of 600,000 won, labor costs of 78,760 won, and material costs of 68,000 won, and can be determined as 59,700 won according to rounding rules (rounding down to the nearest 100 won, etc.).

[0317] In this case, the reserve ratio, additional rate, and application standard table are merely examples, and the actual values ​​may vary depending on the complex's accounting policy, facility risk rating system, or the guidelines of the management agency.

[0318] Through this, the device (200) can calculate reserves by reasonably reflecting prediction uncertainty due to seasonal and operational variables, and improve the stability and transparency of budget management by securing the financial buffering capacity of each complex.

[0319] In step S606, the device (200) can calculate the total amount of accounting items by summing up maintenance costs, labor costs, material costs and reserve costs.

[0320] That is, the device (200) can calculate the total amount of accounting items by aggregating maintenance costs, labor costs, material costs, and reserve costs calculated for each work unit, and accumulating them on a project unit, monthly unit, or complex unit basis.

[0321] Here, the aggregation criteria by item can be established in a hierarchical structure according to work classification (inspection, light maintenance, overhaul, etc.), equipment type (electrical, mechanical, fire protection, sanitation, etc.), and complex area (underground parking lot, rooftop, common hall, etc.).

[0322] When there are multiple tasks within the same complex, the device (200) can calculate not only the total amount per item but also the total amount for the entire complex and the sum of the total amount of the grouped complexes in parallel and transmit them to the upper accounting management unit.

[0323] Additionally, the device (200) may apply VAT, fees, and contract adjustments (fixed amount or percentage basis) as post-processing rules if necessary.

[0324] For example, the device (200) can calculate the final estimated billing amount as “maintenance cost 600,000 won + labor cost 78,760 won + material cost 68,000 won + reserve 59,100 won = total accounting item amount 805,860 won”, then apply a VAT rate of 10% to add VAT of 80,586 won, and reflect an outsourcing contract commission rate of 3% in a separate account.

[0325] At this time, the value-added tax rate, commission rate, adjustment ratio, etc., may be set according to the contract conditions or management regulations for each complex, and this is merely an example and may be set differently depending on the embodiment.

[0326] The device (200) can automatically record the total amount of accounting items and each detailed item calculated in this way into an accounting management database and provide them to the terminal (100) of the apartment building manager in the form of an accounting report classified by item, complex, and period.

[0327] Through this, the device (200) can clearly distinguish the cost structure from work units to complex / group units and prevent omissions or duplicate calculations for each accounting item, thereby ensuring both the accuracy of budget execution and the transparency of financial management.

[0328] In step S607, the device (200) can determine whether the total amount of the calculated accounting item exceeds a preset budget limit by comparing it with the pre-set budget limit, generate budget alarm data if it exceeds it, and if it does not exceed it, reflect the total amount of the accounting item in the grouped management data by complex.

[0329] That is, the device (200) can determine whether the budget is exceeded by querying the pre-set budget limits managed by complex, month, and item (maintenance costs, labor costs, material costs, reserve costs) from the accounting management database and comparing the calculated total amount of accounting items with the budget limits for each item.

[0330] The pre-set budget limits here are values ​​defined in advance based on apartment management regulations, past budget execution history, complex size, degree of facility deterioration, management goals, etc., and may be set differently depending on the characteristics of each complex, for example, “maintenance costs: 15 million won per year (1.25 million won per month), labor costs: 2 million won per month, material costs: 1 million won per month”.

[0331] These budget limits can be configured to be automatically updated by the device (200), and management standards may be adjusted through quarterly or annual verification.

[0332] The device (200) can automatically generate budget alarm data if a budget overrun occurs as a result of comparison.

[0333] This budget alert data includes the excess amount, excess rate, excess items, cause analysis information, recommended measures, and reflection time, and is configured to transmit this in the form of a notification to the terminal (100) of the apartment building manager so that it can be immediately verified and taken action.

[0334] For example, if the total amount of accounting items exceeds the “maintenance budget limit of 1,200,000 won for this month” by 1,265,000 won, the device (200) can generate budget alarm data such as “Excess amount: 65,000 won (Excess rate 5.4%), Cause item: Increase in material costs, Recommendation: Need to order for extended delivery or review alternative materials” and transmit this data in real time to the terminal (100) of the apartment building manager.

[0335] Conversely, if the total amount of an accounting item is below the budget limit, the device (200) can manage the budget execution history by confirming the total amount of the item in the grouped management data by complex and displaying the remaining budget as "carried over to the next month" or "converted to reserve fund."

[0336] Through this, the device (200) can identify the risk of budget overruns by complex at an early stage, support the manager in responding quickly, and immediately reflect the budget within the normal range, thereby increasing the real-time nature and transparency of accounting management.

[0337] Ultimately, the device (200) can prevent budget waste and maximize management efficiency by controlling the financial execution of the apartment complex based on data.

[0338] FIG. 7 is a flowchart illustrating the process of establishing a concentrated patrol zone according to one embodiment.

[0339] Specifically, the device (200) can calculate a risk index for each zone within a multi-unit housing complex to set up a concentrated patrol zone, generate a patrol plan, and reflect the generated patrol plan in a grouped operation plan for each complex.

[0340] That is, the device (200) can integrate the generated grouped patrol plan by complex into the existing grouped operation plan by complex to automatically reflect the work schedule, movement path, and priority allocation time of security personnel.

[0341] For example, the device (200) can be configured to deploy two security personnel to an underground parking lot area classified as a “concentrated patrol area,” shorten the inspection schedule of the common area during daytime hours by 30 minutes according to the patrol plan, and transmit the entire plan to the manager’s terminal (100) so that it can be immediately reviewed and approved.

[0342] Through this, the device (200) can realize an integrated management system that can flexibly respond to security situations in each complex by linking the operation plan adjustment phase and the patrol plan in real time.

[0343] Referring to FIG. 7, first, in step S701, the device (200) can calculate a potential risk index for each zone by checking the past alarm history, crime frequency, and facility failure history for each zone within the apartment complex based on the apartment complex database.

[0344] That is, the device (200) can analyze safety-related historical data for each zone stored in a multi-unit housing database to quantify the patterns of accidents and anomalies occurring over a certain period (e.g., the last 12 months) and calculate a potential risk index indicating a long-term risk level based on this.

[0345] Here, safety-related historical data stored in the multi-unit housing database may consist of alarm occurrence history (fire, intrusion, electric shock, equipment malfunction, etc.), crime occurrence frequency (statistics linked to police report data such as intrusion, theft, and vandalism), and facility failure history (inspection history of lighting, elevators, pumps, etc.).

[0346] Each data can be structured and stored as quantitative indicators such as the number of occurrences, the interval between occurrences, the scale of damage, and the time required for recovery, and the device (200) can process this to be comparable by normalizing it into hourly or monthly units.

[0347] In this case, the pre-set criteria can be defined in the form of a formula of “risk occurrence frequency × weighting factor + failure history × correction factor + alarm history × importance,” where the weighting factor and correction factor may be assigned differently depending on the use of the area (e.g., underground parking lot, rooftop, corridor, etc.), user density, and whether access is restricted.

[0348] For example, if the device (200) detects 4 security alarms (weight 0.5), 3 lighting failures (correction factor 0.3), and 1 outsider intrusion alarm (importance 0.7) in the underground parking lot during the last 12 months, it can calculate a total of 4×0.5 + 3×0.3 + 1×0.7 = 3.4 points, and convert this to a 100-point scale to calculate a potential risk index of 68 points.

[0349] At this time, the weight, correction factor, and importance are merely examples and may be set differently depending on the embodiment.

[0350] Through this, the device (200) can objectively quantify the cumulative risk based on past data and secure zone-specific risk indicators that can be used as reference data for setting up intensive patrol zones in a subsequent stage.

[0351] In step S702, the device (200) can collect real-time video data from CCTVs installed in the apartment complex and extract real-time brightness data for each installation zone of lighting equipment, including streetlights and security lights.

[0352] That is, the device (200) can collect video streams (frame-unit RGB video) of CCTVs installed in each zone of a multi-unit housing complex at regular intervals (e.g., every 5 seconds) and analyze average illuminance values, maximum and minimum illuminance values, contrast by color channel, shadow area ratio, etc. based on pixel brightness distribution in the video to extract real-time brightness data that reflects the lighting conditions of the zone.

[0353] Here, the preset criteria may consist of a 'standard illuminance value for the corresponding time period (e.g., 20 lux at night, 80 lux during the day)' and a 'standard deviation for maintaining lighting (±10%)', wherein the standard illuminance value for each time period may be set differently depending on the type of zone (underground parking lot, pedestrian walkway, park, etc.), the type of lighting fixture (LED, metal halide, etc.), and the installation height.

[0354] For example, the device (200) can determine that the average illuminance value of a specific area during the night time period (22:00~06:00) is 14 lux or less, which is 30% lower than the reference illuminance value of 20 lux, is a ‘lighting failure,’ and determine that the illuminance value is ‘possible lighting failure’ if it drops to 8 lux or less, thereby classifying the brightness grade of the area as “normal / caution / danger.”

[0355] Additionally, the device (200) can detect cases where repeated illumination degradation occurs in the same area in conjunction with lighting maintenance history or power consumption patterns, identify this as a continuous illumination degradation pattern, and assign a risk weight.

[0356] At this time, the reference illuminance value, deviation range, time zone classification, and weighting calculation method are merely examples and may be set differently depending on the embodiment.

[0357] By doing so, the device (200) converts the lighting status of each zone from real-time video data into numerical brightness data, thereby quickly identifying zones with reduced illumination at night, and can reflect them as illumination-based risk factors when calculating the integrated patrol risk index in a subsequent step (S705).

[0358] In step S703, the device (200) can analyze video data collected from the CCTV to detect abnormal events based on preset criteria and generate real-time alarm data.

[0359] That is, the device (200) inputs collected CCTV video frames into an artificial intelligence-based object recognition and behavior analysis model and analyzes the movement trajectory, dwell time, movement pattern, and whether an object (person, vehicle, other object) within each frame has entered an area to automatically detect abnormal events.

[0360] Here, an abnormal event can be defined as a situation where the deviation from normal time, behavior, and location patterns exceeds a certain standard, and the pre-set standard may consist of conditions such as “a person object’s dwell time of 3 minutes or more,” “detection of entry into an unauthorized area,” “a person object that is not moving (falling down),” “a vehicle driving in the wrong direction,” and “access to equipment within a closed area.”

[0361] For example, the device (200) can classify as abnormal events if it detects a person shape that has not moved for more than 5 minutes in a playground area or detects a vehicle that has deviated from a designated movement path in an underground parking lot.

[0362] When an abnormal event is detected, the device (200) can generate real-time alarm data including location information of the area, type of event, time of detection, and risk level (e.g., caution / warning / danger).

[0363] In addition, if the same event is repeated more than a certain number of times in the same area, the device (200) can accumulate and increase the risk level of the area and additionally reflect it when calculating the integrated patrol risk index in the subsequent step (S705).

[0364] At this time, the event detection conditions, judgment thresholds, and repetition accumulation rules are merely examples and may be set differently depending on the embodiment.

[0365] Through this, the device (200) can implement a dynamic anomaly detection system that reflects not only simple static illumination data but also the behavioral patterns of people and vehicles in real time, thereby enabling early detection of real-time dangerous situations within the apartment complex and increasing patrol priority and response efficiency.

[0366] In step S704, the device (200) can calculate an integrated patrol risk index for each zone by integrating potential risk index, real-time brightness data for each zone, and real-time alarm data.

[0367] That is, the device (200) can calculate an integrated patrol risk index by matching potential risk index, lighting brightness level, real-time alarm frequency and severity in the same zone unit and assigning weights according to the importance of each item.

[0368] Here, the potential risk index can be defined as a long-term risk level based on past alarm, failure, and crime history (0 to 100 points), brightness data as a real-time illuminance grade through video analysis (0 to 100 points, higher score for brighter), and real-time alarm data as the frequency and severity of abnormal event detection per unit time (0 to 100 points, higher score for higher frequency and risk).

[0369] The device (200) can set a weighting ratio for each item according to a preset standard to combine these three indicators.

[0370] For example, if the management policy prioritizes the prevention of safety accidents, a weighting ratio of 30% for potential risk index, 30% for brightness data, and 40% for real-time alarm data can be applied.

[0371] At this time, the weighting ratio may be adjusted by considering management regulations, seasonal nighttime activity levels, illuminance standards, recent accident type statistics, etc., and this is merely an example and may be set differently depending on the embodiment.

[0372] For example, the device (200) can receive a potential risk index of 75 points, a brightness score of 60 points, and an alarm score of 95 points for the ‘underground parking lot A’ area and generate an integrated patrol risk index of 80 points calculated as (75×0.3)+(60×0.3)+(95×0.4)=80 points.

[0373] In addition, the device (200) can immediately respond to real-time risk fluctuations by adding a correction factor (+5%) when multiple high-risk alarms occur within a short period in a specific area (e.g., 3 or more / 1 hour).

[0374] Through this, the device (200) can obtain a quantitative patrol priority indicator that comprehensively reflects long-term risks, environmental illuminance, and real-time anomaly detection results, and can reliably provide base data for setting up intensive patrol zones.

[0375] In step S705, the device (200) can set an area where the integrated patrol risk indicator is greater than or equal to a preset threshold value as a concentrated patrol area.

[0376] That is, the device (200) can compare the calculated integrated patrol risk index of each zone with a reference value and automatically classify zones where the index is greater than or equal to the reference as intensive patrol zones.

[0377] Here, an intensive patrol zone refers to an area with a high complex risk of potential dangers, reduced illumination, and real-time abnormal events, requiring priority personnel deployment and increased patrol frequency.

[0378] The device (200) can set a preset standard value by considering complex factors such as management policy, security grade of the complex, number of night patrol personnel, local crime rate, and seasonal accident frequency.

[0379] For example, the device (200) may define a standard value as “integrated patrol risk index 80 points or more: intensive patrol required area”, “60 to 79 points: general patrol area”, and “59 points or less: low risk area”, and this is merely an example and the weight, score range, and classification section may be set differently depending on the embodiment.

[0380] Additionally, the device (200) may apply supplementary rules to automatically designate the top 20% as intensive patrol zones by considering the relative distribution of risk levels by complex when setting the standard value, or exceptionally include zones where real-time alarms have occurred continuously within the last 7 days as intensive patrol targets.

[0381] For example, the device (200) can set 'underground parking lot A (risk index 86 points)' and 'elevator hall C (risk index 84 points)' as intensive patrol areas during nighttime hours, and classify 'playground B (risk index 58 points)' as a general patrol area.

[0382] In addition, since the influence of light intensity is minimal during the daytime, the threshold value can be relaxed by 5 points (e.g., intensive patrol zones with 75 points or more) and applied dynamically depending on the time of day.

[0383] Through this, the device (200) can prevent the dispersed deployment of security resources and maximize patrol efficiency centered on high-risk areas by automatically designating a concentrated patrol area that takes into account time, location, and environmental conditions.

[0384] In addition, the intensive patrol zone information is utilized as input values ​​for calculating the number of patrols, routes, and times in a subsequent step (S706), providing a basis for automatically generating customized patrol plans for each complex.

[0385] In step S706, the device (200) can generate a grouped patrol plan by calculating the number of patrols, the patrol route, and the patrol time according to preset criteria for the concentrated patrol area.

[0386] That is, the device (200) can comprehensively analyze the integrated patrol risk indicators, location, accessibility, area of ​​each zone, crime statistics by time of day, CCTV field of view overlap, etc., for each zone of the concentrated patrol zone, and accordingly, automatically calculate the number of patrols, patrol routes, and patrol times according to preset standards.

[0387] Here, the pre-set criteria serve as judgment criteria for determining patrol frequency, route planning, and patrol time intervals, and can be defined as, for example, “risk index 90 points or higher: patrol 6 times or more per day (at intervals of 2 to 3 hours), 80 to 89 points: patrol 4 times (at intervals of 4 hours), 60 to 79 points: patrol 2 times (at intervals of 6 hours), 59 points or lower: patrol 1 time (at intervals of 8 hours).”

[0388] Additionally, the device (200) can calculate a route according to the principle of “minimizing travel distance between identical zones and maximizing the inclusion ratio of blind spots (zones with illuminance of 10 lux or less) in the patrol route,” and can simultaneously apply the condition of “strengthening intensive night patrols (shortening patrol intervals, prioritizing placement within the route) in zones with illuminance values ​​of 15 lux or less.”

[0389] At this time, these standards are merely examples and may be set differently depending on the embodiment, merely according to scale, number of personnel, CCTV coverage ratio, safety management regulations, etc.

[0390] For example, the device (200) may designate the ‘underground parking lot A (risk index 86 points)’ area as a patrol target 5 times a day, set the ‘security light area B (illumination 10 lux)’ as a night intensive patrol target 3 times a day, and designate the ‘playground C (risk index 62 points)’ as a patrol target 2 times a day.

[0391] Additionally, when the device (200) automatically calculates the movement path between patrol zones, it can calculate the path based on the shortest distance (e.g., 200m) in the order A→B→C by considering the average movement distance between zones, the starting position of the security personnel's patrol, restricted passage sections, and overlapping CCTV view sections.

[0392] The device (200) can be configured to automatically generate a patrol plan table by complex and time zone by integrating the calculated number of patrols, routes, and patrol times, and to visually display this on the terminal (100) of the apartment building manager so that the manager can review and approve the deployment of patrol personnel and the schedule.

[0393] Additionally, the device (200) can store the generated patrol plan in an operational database and subsequently accumulate and manage feedback data that analyzes the performance rate and execution efficiency relative to the plan by comparing it with the actual patrol results (e.g., patrol logs, video verification results).

[0394] Through this, the device (200) can automatically calculate the risk-based patrol frequency, route, and time to prevent unnecessary duplicate patrols, implement a concentrated monitoring system centered on high-risk areas, and maximize the efficiency of plan verification and performance management through the manager's terminal (100).

[0395] FIG. 8 is a flowchart illustrating the process of predicting the life cycle of equipment and generating replacement and maintenance schedules according to one embodiment.

[0396] Specifically, the device (200) can predict the life cycle of equipment by grouped complex based on equipment condition prediction information and generate replacement and maintenance schedules.

[0397] Referring to FIG. 8, first, in step S801, the device (200) can calculate the cumulative operating time and average load rate for each facility based on operating time data, load rate data, and environmental variable data collected through IoT sensors installed in each facility within the apartment complex.

[0398] That is, the device (200) can collect operation-related data such as operating time, load rate, temperature, humidity, voltage fluctuation, vibration, etc. from IoT sensors installed in major facilities (e.g., boiler, pump, ventilation fan, elevator, etc.) within a multi-unit housing complex at regular intervals, and calculate the cumulative operating time and average load rate for each facility after applying a data validity period (e.g., the last 3 months or 6 months), measurement interval (e.g., 1 minute, 10 minutes, 1 hour), and missing value correction rules (e.g., linear interpolation or keeping recent values) according to preset criteria.

[0399] Here, operating time data can be defined as the sum of the equipment's power supply time, actual operation duration, and standby time, and load rate data can be defined as the average ratio of the real-time load to the equipment's rated capacity.

[0400] In addition, environmental variable data may include external conditions such as temperature, humidity, vibration intensity, power quality, and air quality of the space where the equipment is located.

[0401] For example, the device (200) can calculate the actual operating time converted value as 3,680 hours by applying a load correction factor (e.g., +15% when the load rate exceeds 65%) based on data where the cumulative operating time of the ventilation fan is 3,200 hours, the average load rate is 68%, and the average ambient temperature is maintained at 40℃.

[0402] At this time, the preset criteria may be set to include an influence coefficient of the load rate, an environmental temperature correction coefficient, a data aggregation unit, etc., and this is merely an example and may be set differently depending on the embodiment.

[0403] Through this, the device (200) can build a highly reliable operation history database that reflects actual operating conditions and load effects, rather than a simple sum of accumulated time.

[0404] This data can subsequently be used as foundational data for estimating remaining lifespan and predicting replacement and maintenance schedules.

[0405] In step S802, the device (200) can calculate the remaining lifespan of each facility according to a preset standard by comparing the cumulative operating time, average load rate, and environmental variable data with a preset standard lifespan value for each facility.

[0406] That is, the device (200) can calculate the degree of lifespan consumption under actual usage conditions by referring to the calculated cumulative operating time and average load rate and environmental variable data for each facility and comparing it with the standard lifespan value corresponding to each facility type.

[0407] Here, the predefined standard life values ​​for each piece of equipment can be derived based on life data provided by the manufacturer, durability standards defined in national standards (KS, ISO, etc.), or maintenance history statistics of the same group of equipment, for example, the boiler can be set to 10,000 hours, the pump to 8,000 hours, and the elevator motor to 12,000 hours.

[0408] In addition, the preset criteria include life correction rules based on actual operating conditions, for example, if the operating time is 80% or more of the standard life, an elapsed correction factor (e.g., 1.05) is applied, if the average load rate exceeds 70%, an accelerated degradation factor (e.g., 1.10) is applied, and if the ambient temperature is 35℃ or higher, an environmental shortening rate (e.g., 5%) may be additionally reflected.

[0409] These standards may be set differently depending on the operating environment of the equipment, material characteristics, cooling method, etc., and are merely examples and may be set differently depending on the embodiment.

[0410] For example, the device (200) can check the cumulative operating time of 6,400 hours (80%) relative to the standard lifespan of 8,000 hours of the pump, the average load rate of 75%, and the average temperature of 38°C, and apply an accelerated deterioration factor of 1.1 and an environmental reduction rate of 5% to calculate the lifespan consumption as 6,400 × 1.1 × 1.05 = 7,392 hours and calculate the remaining lifespan as approximately 608 hours.

[0411] Through this, the device (200) can secure a quantitative remaining life assessment system that comprehensively considers the actual operating environment and load conditions, rather than a simple usage time basis.

[0412] In step S803, the device (200) can generate early warning data for equipment with a high probability of early failure by comparing the remaining lifespan calculation results with past failure history data and manufacturer warranty period information.

[0413] That is, the device (200) can determine equipment that is likely to fail earlier than its normal lifespan and generate early warning data for it by analyzing the equipment's past failure history data (including the time of failure, cause, and parts replacement history) and manufacturer's warranty period information (including information on the period during which normal operation is guaranteed and the recommended replacement cycle for parts) together, based on the remaining lifespan of each piece of equipment.

[0414] Here, remaining life refers to the remaining usable time or cycle calculated based on the cumulative operating time, average load rate, and environmental variable data of the equipment, and failure history data refers to a record including the past failure frequency of the equipment, whether the same part was repeatedly replaced, and the time to failure (Mean Time To Failure, hereinafter referred to as MTTF).

[0415] In addition, manufacturer warranty information indicates the normal usage warranty period or the recommended replacement cycle for major components provided by the manufacturer, and early warning data is information generated to identify equipment with a high probability of early failure in advance, and may include equipment ID, risk level, expected time of failure, and type of recommended action (e.g., inspection, parts replacement, performance diagnosis).

[0416] The device (200) can quantify the risk of early failure for each piece of equipment by comprehensively analyzing the repeat replacement pattern of the same part or the remaining life level, including cases where the MTTF is less than 50% of the standard value, by evaluating the failure interval variability of each piece of equipment.

[0417] The judgment criteria may be set such that if one or more of the following are satisfied, it is determined that there is a high probability of early failure: “the remaining life is 20% or less of the standard life,” “two or more failures of the same type occur before the end of the warranty period,” or “MTTF is less than 50% of the standard value,” these conditions may be variably adjusted according to the importance and safety impact of the equipment.

[0418] For example, if the standard life of the pump equipment is 8,000 hours, the device (200) can generate early warning data in the condition of recommending preventive inspection by classifying the equipment as “high probability of early failure” when the remaining life is only 1,200 hours (15%) and there is a record of the same bearing defect occurring two or more times within the warranty period (3 years).

[0419] In addition, the device (200) can transmit the generated early warning data to the terminal (100) of the apartment building manager to support the manager in immediately identifying equipment at risk of failure and taking preventive measures.

[0420] Through this, the device (200) can establish a reliability-based early warning system that integrates actual past failure patterns and manufacturer standards, going beyond simple lifespan calculation, and can improve the operational stability of the facilities in the apartment complex by preventing unexpected facility shutdowns or safety accidents in advance.

[0421] In step S804, the device (200) can calculate the failure probability based on failure history data of similar equipment types and calculate the equipment replacement priority using the failure probability and remaining life data.

[0422] That is, the device (200) can statistically analyze past failure history data belonging to the same type of equipment (e.g., boiler, pump, elevator, ventilation fan, etc.) to derive a failure probability function based on the passage of time or operating time, and calculate a replacement priority by reflecting the remaining lifespan data for each piece of equipment.

[0423] At this time, the device (200) can analyze the failure time, type of cause, and recurrence cycle after repair from the failure history data to calculate the cumulative failure rate, mean time between failures (MTBF), and failure probability distribution (e.g., Weibull distribution, log-normal distribution) for each piece of equipment.

[0424] In addition, the pre-set criteria can be defined in the form of “replacement priority score = failure probability × (1 remaining life ratio) × importance weight”, where the importance weight can be adjusted within the range of 0.8 to 1.2 depending on the functional importance of the facility (e.g., whether it is a safety, convenience, or essential facility) or the impact within the complex (e.g., number of households, proportion of common areas, etc.).

[0425] For example, the device (200) can calculate the replacement priority score of equipment with “average failure probability of pumps 0.35, remaining life ratio 0.2, importance weight 1.0” as 0.35×(10.2)×1.0=0.28 and classify it as a higher rank than the score of equipment with “failure probability of ventilation fans 0.20, remaining life ratio 0.5, importance weight 1.1” as 0.20×(10.5)×1.1=0.11.

[0426] In addition, if there are facilities with the same priority score, the device (200) can finely adjust the detailed priority by applying auxiliary criteria such as the number of failures, manufacturer grade, and difficulty of obtaining parts.

[0427] Through this, the device (200) can establish an objective replacement priority system that reflects actual failure patterns and equipment importance, rather than simply judging based on remaining lifespan alone, and can improve the efficiency and stability of the maintenance plan by allocating budget, materials, and personnel to high-risk equipment based on these results.

[0428] In step S805, the device (200) can determine the timing of equipment replacement and maintenance cycles according to preset criteria based on replacement priority and remaining lifespan per piece of equipment.

[0429] That is, the device (200) can automatically calculate the replacement time (immediate, short-term, medium-term, long-term) and maintenance cycle (regular inspection interval, parts replacement cycle, etc.) for each piece of equipment by comprehensively analyzing the calculated equipment replacement priority score and the remaining lifespan data of each piece of equipment.

[0430] In this case, the pre-set criteria may consist of multi-layered rules that consider the operating characteristics, importance, and environmental impact factors of the equipment.

[0431] For example, criteria such as “remaining life ≤ 500 hours: subject to immediate replacement”, “500 to 1500 hours: plan to replace within 1 month”, “1500 to 3000 hours: shortening of regular maintenance cycle (existing 30 days → 20 days)”, and “3000 hours or more: maintaining existing cycle” may be set.

[0432] Additionally, the device (200) can set the replacement time differently even with the same remaining lifespan according to the importance weight of the equipment.

[0433] For example, critical safety equipment such as fire pumps can be classified as 'short-term (within 2 weeks) replacement' when the remaining lifespan is 1,000 hours, while non-critical equipment such as ventilation fans can be set as 'medium-term (within 1 to 2 months) replacement' even under the same conditions.

[0434] In addition, the device (200) can reflect the load rate and environmental conditions together when calculating the maintenance cycle.

[0435] For example, in the case of high-load equipment with an average load rate of 80% or more or equipment operating in an environment with an ambient temperature of 40°C or higher, the existing inspection cycle of 30 days can be automatically corrected to be shortened to 15 to 20 days.

[0436] Conversely, for equipment with a low load and operating within a normal temperature range (25℃ or lower), the efficiency of management resources can be increased by maintaining or relaxing the inspection cycle (e.g., 30 days → 40 days).

[0437] Through this, the device (200) can automatically derive a practical replacement and maintenance execution schedule that reflects the importance, load conditions, and operating environment of each facility, going beyond simple remaining life prediction results.

[0438] As a result, by reducing unnecessary early replacements while strengthening the proactive response capability for high-risk facilities, it is possible to simultaneously secure the stability and efficiency of facility maintenance within apartment complexes.

[0439] In step S806, the device (200) can collect performance data through IoT sensors and calculate a performance deviation index by comparing it with a reference performance value.

[0440] That is, the device (200) can periodically collect key performance data such as temperature, voltage, pressure, vibration, and efficiency from IoT sensors attached to each facility, and calculate a performance deviation index by comparing the data with a standard performance value for each facility.

[0441] Here, the reference performance value can be set as the average value of data measured at the time of the initial normal operation of the equipment, specified specifications provided by the manufacturer (rated efficiency, standard operating pressure, etc.), or the statistical median of long-term steady-state data.

[0442] At this time, the preset standard can be defined by the formula “Deviation rate = (Actual value standard value) / Standard value × 100”, and the condition of the equipment can be divided into three stages based on the calculated deviation rate: “Normal (deviation rate ≤ ±10%)”, “Caution (±10~25%)”, and “Abnormal (±25%)”.

[0443] For example, the device (200) can store a performance deviation index of 16 by applying a deviation rate calculated as (4.25.0) / 5.0×100 = 16% when the reference pressure value of the pump is 5.0 bar and the real-time sensor value is 4.2 bar, and classify this as a 'caution section'.

[0444] Additionally, the device (200) can calculate a deviation index for each of the various performance items (e.g., temperature, voltage, vibration, etc.) of the same equipment, and then calculate an integrated performance deviation index through a weighted average.

[0445] For example, considering the items of pressure (weight 0.4), temperature (0.3), and vibration (0.3), the integrated performance deviation index calculated as (16×0.4)+(8×0.3)+(5×0.3)=9.7 can be determined as “minor caution.”

[0446] At this time, the weights, intervals, and judgment criteria are merely examples and may be set differently depending on the embodiment.

[0447] For example, for some high-precision equipment (e.g., elevator control modules), the setting may be tightened to 'caution' when the deviation rate exceeds 5%, and conversely, for equipment with a large environmental impact (e.g., ventilation fans), it may be allowed to be within the 'normal' range of ±20%.

[0448] Through this, the device (200) can go beyond a simple threshold detection level and quantify the degree of performance degradation of the equipment to more precisely predict when maintenance is needed and the priority of preventive maintenance.

[0449] Consequently, the calculation of a performance deviation index based on real-time performance data can serve as a basis for the device (200) to identify the deterioration trend of the equipment early and improve the accuracy of failure prediction.

[0450] In step S807, if the performance deviation index exceeds a preset reference value, the device (200) can automatically recommend a response manual by referring to preset failure cause data by equipment type.

[0451] That is, when the calculated performance deviation index exceeds a reference value, the device (200) can identify a possible cause of the problem by referring to a fault cause database corresponding to the type, structure, and usage environment of the equipment, and automatically recommend a response manual including a maintenance procedure or inspection guide.

[0452] Here, the preset reference values ​​may be defined differently for each piece of equipment, taking into account the type of equipment, service life, operating environment, load pattern, etc.

[0453] For example, it can be set to ±25% for boilers operating in high-temperature environments, ±15% for motors and pumps with high vibration sensitivity, and ±10% for electrical control equipment (e.g., elevator control panels).

[0454] Additionally, the device (200) can automatically classify major abnormal items of the equipment (e.g., temperature rise, vibration increase, efficiency decrease, etc.) when the performance deviation index exceeds a reference value, and can call up fault cause data that matches these items.

[0455] For example, if the pump's performance deviation index exceeds the reference value (±20%) by +24%, the device (200) can refer to a failure cause rule corresponding to the equipment type in an internal database to derive possible causes such as "possible packing wear," "possible bearing deterioration," and "possible misalignment of the rotating shaft," and automatically recommend a maintenance manual for these causes.

[0456] The recommended manual may include inspection items (e.g., checking packing condition, checking bearing replacement cycle), a list of necessary materials (e.g., packing set, lubricant), and estimated inspection time.

[0457] In this case, failure cause data can be configured based on past maintenance history, manufacturer-provided manuals, and real-time sensor pattern analysis results, and can be managed in the form of cause-action mapping tables for each equipment type.

[0458] For example, probability-based mapping is possible, such as “increased vibration → bearing damage (probability 0.65), rotational axis imbalance (probability 0.25), pump housing loosening (probability 0.10)”.

[0459] Through this, the device (200) can realize a cause-centered maintenance response system beyond simple alarm generation, and can enable preemptive response through the terminal (100) of the apartment building manager by automating the execution steps of preventive maintenance or predictive maintenance.

[0460] As a result, the device (200) can improve the pre-failure response rate and minimize operational losses due to unexpected shutdowns by automatically recommending a rapid and evidence-based response manual for equipment where the performance deviation index exceeds the standard.

[0461] In step S808, the device (200) can calculate a risk level for each piece of equipment by integrating information on remaining lifespan, replacement priority, maintenance cycle, performance deviation index, and response manual, predict the life cycle of the equipment based on the risk level and maintenance cycle, and generate a replacement and maintenance schedule according to the predicted life cycle and risk level.

[0462] That is, the device (200) can calculate the risk level for each facility by integrating and analyzing the key indicators of each facility.

[0463] The device (200) can predict the life cycle of each facility based on these integrated analysis results and automatically generate a replacement or maintenance schedule according to the risk level.

[0464] Here, the pre-set criteria can be defined as a weighted formula to reflect the condition of the equipment in a multidimensional way, and can be set in the form of, for example, “Risk level calculation formula = (1 remaining life ratio) × 0.4 + (performance deviation index normalized value) × 0.3 + (failure probability) × 0.3”.

[0465] At this time, each item is normalized to a range of 0 to 1 and calculated, and the result is converted to a 100-point scale and can be classified into risk levels “High (80 points or more) / Medium (60 to 79 points) / Low (59 points or less)”.

[0466] For example, if the device (200) measures the remaining life ratio of a specific pump as 0.2, the performance deviation index as 18 (normalized 0.18), and the failure probability as 0.35, (10.2)×0.4 + 0.18×0.3 + 0.35×0.3 = 0.32 + 0.054 + 0.105 = 0.479 → convert this to a 100-point scale to calculate a risk level of 82 points, and classify the equipment as “high risk level.”

[0467] Afterward, the device (200) can generate a replacement schedule within 7 days for equipment with a risk level of “high,” a repair plan within 30 days for equipment with a risk level of “medium,” and a maintenance schedule at the level of regular inspection for equipment with a risk level of “low.”

[0468] In addition, when generating a replacement and maintenance schedule, the device (200) can automatically adjust the schedule to an actually feasible schedule by taking into account the location of the equipment, available personnel, material availability, and dependency relationship with other equipment (e.g., range of impact when equipment within the same line is stopped).

[0469] For example, the device (200) can set “Pump A (risk level 82 points)” as a replacement target within 7 days and automatically classify “Boiler B (risk level 68 points)” as a maintenance target in the next inspection cycle (after 20 days), and when creating a schedule, it can simultaneously link personnel allocation and material ordering plans to display as an integrated schedule on the manager’s terminal (100).

[0470] Through this, the device (200) can implement a risk-based lifecycle prediction and maintenance schedule automation system beyond simple equipment condition diagnosis.

[0471] As a result, the device (200) can simultaneously secure the efficiency of maintenance and the accuracy of predictive maintenance by integrating equipment-specific status data, failure history, and environmental factors to predict replacement and repair timing, and by realizing an intelligent equipment management schedule generation function based on this.

[0472] FIG. 9 is a flowchart illustrating the process of generating a performance report based on grouped management data by complex according to one embodiment.

[0473] Specifically, the device (200) can analyze operational efficiency between grouped complexes based on grouped complex-specific management data and generate a performance report based on the analysis results.

[0474] Referring to FIG. 9, first, in step S901, the device (200) can extract maintenance cost, labor cost, and material cost data from grouped complex-specific accounting management information and calculate the cost reduction rate for each complex by comparing it with the average value for each complex within the group.

[0475] That is, the device (200) can extract data on maintenance costs, labor costs, and material costs for each complex stored in the grouped accounting management information for each complex, and calculate the cost reduction rate for each complex by comparing it with the average value of other complexes within the same group.

[0476] At this time, the comparison criteria for calculating the cost reduction rate may include a formula and judgment rules for quantitatively evaluating the cost efficiency of each complex within the group, and may be defined, for example, in the form of “cost reduction rate = (group average cost / corresponding complex cost) ÷ group average cost × 100”.

[0477] In addition, if the reduction rate is 0 or greater, it can be determined as a reduction, and if it is less than 0, it can be determined as an overspending.

[0478] Here, grouped accounting management information by complex refers to a data set for integrated management of accounting items (maintenance costs, labor costs, material costs) of multiple complexes belonging to the same management group.

[0479] Maintenance costs refer to direct costs incurred for equipment inspection, repair, and replacement; labor costs refer to the cost of manpower input based on the work schedules of management personnel; and material costs refer to costs calculated based on the unit price and quantity of ordered and delivered materials.

[0480] In addition, the cost reduction rate per complex refers to an indicator that expresses the degree of cost reduction of a specific complex relative to the group average as a percentage.

[0481] For example, if the maintenance cost of 'A complex' is 9.2 million won and the group average is 10 million won, the device (200) calculates the reduction rate as (1,000920) / 1,000×100=8%, and combines the labor cost reduction rate of 5% and the material cost reduction rate of 3% to calculate the total cost reduction rate of 5.3%.

[0482] Through this, the device (200) can quantitatively evaluate the budget execution efficiency of each complex and identify complexes within the group where cost reduction is insufficient to derive additional management improvement directions.

[0483] In step S902, the device (200) can extract work schedule and personnel placement data from grouped complex-specific operation plan information and calculate a personnel input efficiency index according to preset criteria.

[0484] That is, the device (200) can analyze grouped operation plan information for each complex to check the work schedule, personnel deployment status, and amount of work performed relative to working hours for each complex, and calculate the personnel input efficiency index.

[0485] At this time, the pre-set criteria may include a calculation formula and an evaluation grade system to quantify work efficiency relative to manpower input, and may be defined, for example, in the form of 'Manpower Input Efficiency Index = Number of Actual Work Completed ÷ Input Working Hours × 100'. Additionally, a standard value of 1.0 may be set as average efficiency, and 1.2 or higher may be judged as high efficiency, less than 1.0 to 0.8 or higher as average, and less than 0.8 as inefficient.

[0486] Here, grouped operational plan information by complex refers to integrated management data including work schedules, work zones, personnel input, and work types for each complex belonging to the same group.

[0487] Work schedule refers to data indicating each employee's work time slot (day, night), shift cycle, and work location, while personnel deployment data refers to information indicating how personnel are assigned by zone or work type within the complex.

[0488] In addition, the manpower input efficiency index refers to an indicator that quantifies manpower management efficiency by converting the amount of work actually performed relative to the working hours of personnel deployed over a certain period into a ratio.

[0489] For example, if the device (200) has security, cleaning, and facility management personnel from ‘B Complex’ deployed 24 hours a day to complete 30 inspection and maintenance tasks, the manpower input efficiency index is calculated as 30 / 24=1.25, which can be evaluated as showing an efficiency 25% higher than the average.

[0490] Through this, the device (200) can quantitatively evaluate the efficiency of manpower management by complex and identify areas where manpower placement is excessive or insufficient, thereby providing basis data for future work schedule adjustments and manpower redeployment plans.

[0491] In step S903, the device (200) can check material unit price and delivery date data from the order information and material management database of materials by grouped complex and calculate a material procurement efficiency index by complex by comparing with the average value within the group.

[0492] That is, the device (200) can verify the unit price and delivery date data of materials for each group by linking the order information of materials for each group with the material management database, and calculate the material procurement efficiency index by comparing it with the group average value.

[0493] At this time, the comparison criteria for calculating the material procurement efficiency index may include a formula and a grading system for evaluating material unit price and delivery time efficiency in combination, and may be defined, for example, in the form of “Material Procurement Efficiency Index = (Unit Price Efficiency Score + Delivery Time Efficiency Score) ÷ 2”.

[0494] Unit price efficiency can be calculated as 'group average unit price ÷ unit price of the complex × 100', and delivery time efficiency as 'group average delivery date ÷ delivery date of the complex × 100', and an efficiency index of 100 points or more can be judged as high efficiency, 90 to 99 points as average, and less than 90 points as inefficient.

[0495] Here, the order information for grouped materials by complex refers to data including the required material items, specifications, order quantities, delivery dates, and supplier information for each complex, and the material management database refers to an internal database storing contract unit prices, delivery history, supplier reliability, delivery delay history, etc.

[0496] Material unit prices refer to the contract or recent actual transaction amounts per unit item, and delivery time data refers to the average time required from the time of order placement to the completion of delivery.

[0497] In addition, the material procurement efficiency index is a comprehensive indicator calculated by combining efficiency in terms of unit price and delivery time, and it is a value that quantifies the relative efficiency of the material procurement system for each complex.

[0498] For example, the device (200) can calculate a material procurement efficiency index of 107.5 points by averaging the unit price efficiency of 105 points and the delivery time efficiency of 110 points when the unit price of the material of the ‘C complex’ is 5% lower than the group average and the delivery time is 10% faster than the average.

[0499] Through this, the device (200) can quantitatively evaluate the material procurement efficiency by complex to identify bottlenecks in the supply chain and provide basis data that can be reflected in the establishment of future material purchasing policies and supplier evaluations.

[0500] In step S904, the device (200) can calculate the overall operational efficiency index of the grouped complex by combining the cost reduction rate, the manpower input efficiency index, and the material procurement efficiency index.

[0501] That is, the device (200) can calculate a comprehensive operational efficiency index for each complex by combining the three calculated indicators based on weights according to their respective management impacts.

[0502] At this time, the criteria for calculating weights can be set by reflecting the impact of each indicator on operational efficiency. For example, the cost reduction rate may be given a weight of 0.4 because it has a high direct impact on budget execution, while the manpower input efficiency index and the material procurement efficiency index may each be given a weight of 0.3.

[0503] In addition, the calculated overall operational efficiency index can be classified as high efficiency if it is 90 points or higher, average if it is 70 to 89 points, and a range requiring improvement if it is less than 70 points.

[0504] Here, the cost reduction rate refers to the ratio of savings in maintenance, labor, and material costs per complex calculated from accounting management information for each grouped complex, and the manpower input efficiency index refers to the ratio of the actual amount of work performed to the work schedule calculated from operational planning information for each grouped complex.

[0505] The material procurement efficiency index refers to material unit cost and delivery efficiency calculated based on material ordering information and material management databases for each grouped complex, while the overall operational efficiency index refers to a comprehensive indicator that quantifies the overall operational efficiency of the complex by integrating these three indicators.

[0506] For example, the device (200) can calculate (7×0.4)+(115×0.3)+(108×0.3)=83.7 points by substituting the cost reduction rate of 'D Complex', manpower input efficiency index of 1.15 (115 points), and material procurement efficiency index of 108 points, and record this as the overall operational efficiency index.

[0507] Through this, the device (200) can comprehensively diagnose the balance of cost, manpower, and material operations for each complex, and for complexes with low efficiency, it can automatically generate data requiring improvement and reflect this in the adjustment of operation plans and decision-making for resource reallocation.

[0508] In step S905, the device (200) can determine whether efficiency is improved by comparing the overall operational efficiency index with a preset reference value, and if it falls below the reference value, it can generate data requiring improvement.

[0509] That is, the device (200) can determine whether efficiency is improved by comparing the calculated overall operational efficiency index with a preset standard value, and if it falls below the standard, it can automatically identify vulnerable items of the complex and generate data requiring improvement.

[0510] In this case, the preset standard value may be set by referring to a certain percentage of the group average, and this may vary depending on operating conditions such as scale, budget level, workforce composition ratio, and frequency of material usage.

[0511] For example, 90% or more of the group average can be defined as 'good' that meets target efficiency, 70–89% as 'average', and less than 70% as 'needs improvement' that requires management improvement.

[0512] Here, data requiring improvement refers to a set of information designed to automatically identify the causes of inefficiency for each complex and suggest directions for improvement.

[0513] The data may include fields such as identifiers, names of inefficient items, excess rates, results of cause analysis, and recommended improvement measures (e.g., changing suppliers, shortening inspection cycles, reallocating personnel).

[0514] For example, if the device (200) shows that the overall operational efficiency index of ‘E Complex’ is 68 points compared to the group average (85 points), it can analyze the cause in the cost reduction rate item as material costs being 15% higher than the average and generate data requiring improvement including “review of material unit price negotiation” and “recommendation to change suppliers.”

[0515] Through this, the device (200) can automatically diagnose operational inefficiencies in each complex and provide specific improvement directions to the manager's terminal (100) based on data requiring improvement, thereby simultaneously realizing real-time feedback on operational efficiency and shortening the improvement cycle.

[0516] In step S906, the device (200) can generate a performance report including an overall operational efficiency index, a cost reduction rate, a manpower input efficiency index, and a material procurement efficiency index, and transmit the generated performance report to the terminal of the apartment building manager.

[0517] That is, the device (200) can automatically generate a performance report that comprehensively evaluates the performance of each complex by integrating the overall operational efficiency index, cost reduction rate, manpower input efficiency index, and material procurement efficiency index calculated from the grouped complex-specific management data.

[0518] Here, the performance report is a data object that quantitatively summarizes and visualizes key operational indicators for each complex, and may include information such as complex identifiers, operational efficiency grades, indicators by item (maintenance cost reduction rate, manpower input efficiency, material procurement efficiency, etc.), deviations from the group average, items recommended for improvement, and monthly or quarterly trend graphs.

[0519] In addition, performance reports can be structured in the form of tables, graphs, heatmaps, and comparison charts, and can intuitively display rankings by complex, key strengths, and budget execution trends.

[0520] For example, the device (200) can generate a performance report including summary information such as “E complex: operational efficiency 68 points (needs improvement), major improvement item: material cost reduction”, “A complex: operational efficiency 92 points (excellent), major strength: workforce allocation optimization, top 10% material procurement efficiency”, and transmit it to the terminal (100) of the apartment building manager to display it in real time on a dashboard screen or in the form of a monthly report.

[0521] Through this, the device (200) can integrately analyze and visualize the operational efficiency and cost execution results of each complex, thereby enabling the manager to grasp operational performance at a glance at the group level and immediately plan improvement measures for inefficient complexes.

[0522] In addition, by promoting the optimization of budget planning, workforce deployment strategies, and material procurement policies based on automatically generated performance reports, the level of integrated management and operational efficiency of multi-unit housing complexes can be systematically improved.

[0523] FIG. 10 is an example diagram of the configuration of a device according to one embodiment.

[0524] A device (200) according to one embodiment includes a processor (210) and a memory (220). A device (200) according to one embodiment may be the server or terminal described above. The processor (210) may include at least one device described above through FIGS. 1 to 9 or perform at least one method described above through FIGS. 1 to 9. The memory (220) may store information related to the method described above or store a program in which the method described above is implemented. The memory (220) may be volatile memory or non-volatile memory.

[0525] The processor (210) can execute a program and control the device (200). The code of the program executed by the processor (210) can be stored in memory (220). The device (200) can be connected to an external device (e.g., a personal computer or a network) through an input / output device (not shown in the drawing) and exchange data.

[0526] The embodiments described above may be implemented as hardware components, software components, and / or combinations of hardware and software components. For example, the devices, methods, and components described in the embodiments may be implemented using one or more general-purpose or special-purpose computers, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable gate array (FPGA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing unit may execute an operating system (OS) and one or more software applications executed on said operating system. Additionally, the processing unit may access, store, manipulate, process, and generate data in response to the execution of the software. For ease of understanding, the processing unit may be described as being used as a single unit, but those skilled in the art will understand that the processing unit may include multiple processing elements and / or multiple types of processing elements. For example, the processing unit may include multiple processors or one processor and one controller. Additionally, other processing configurations, such as parallel processors, are also possible.

[0527] The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., either alone or in combination. The program instructions recorded on the medium may be those specifically designed and configured for the embodiment, or they may be those known and available to those skilled in the art of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory. Examples of program instructions include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc. The hardware devices described above may be configured to operate as one or more software modules to perform the operation of the embodiment, and vice versa.

[0528] Software may include computer programs, code, instructions, or a combination of one or more of these, and may configure a processing unit to operate as desired or command the processing unit independently or collectively. Software and / or data may be permanently or temporarily embodied in any type of machine, component, physical device, virtual equipment, computer storage medium or device, or transmitted signal wave so as to be interpreted by the processing unit or to provide instructions or data to the processing unit. Software may be distributed over networked computer systems and may be stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

[0529] Although the embodiments have been described above with reference to the limited drawings, those skilled in the art can apply various technical modifications and variations based on the above. For example, suitable results may be achieved even if the described techniques are performed in a different order than described, and / or if the components of the described system, structure, device, circuit, etc. are combined or assembled in a form different from described, or replaced or substituted by other components or equivalents.

[0530] Therefore, other implementations, other embodiments, and equivalents to the claims also fall within the scope of the claims set forth below. Explanation of the symbols

[0532] 100: Apartment building manager's terminal 200: Device 210: Processor 220: Memory

Claims

Claim 1 A method for predicting the condition of multi-unit housing facilities based on artificial intelligence and IoT sensors, and for integrated management of accounting and material-linked operations by complex, performed by a device, comprising: a step of collecting data by multi-unit housing complex based on a multi-unit housing database and grouping it according to preset criteria; a step of collecting facility status data through IoT sensors installed on each facility within the grouped multi-unit housing complex; a step of generating facility status prediction information including the probability of abnormality occurring in each facility and the timing of maintenance requirements by inputting the facility status data into an artificial intelligence model pre-trained based on past facility failure history data and inspection result data; a step of generating ordering information for materials required within a preset period by referring to the grouped complex-specific inventory information stored in a material management database based on the facility status prediction information; a step of adjusting the grouped complex-specific operation plan, including security schedules, cleaning schedules, and common area management schedules, according to preset criteria based on the facility status prediction information; and a step of automatically generating grouped complex-specific accounting management information including maintenance costs, labor costs, material costs, and contingency reserves based on the facility status prediction information and the grouped complex-specific operation plan. and the step of integrating, storing, and managing grouped complex-specific management data including facility status prediction, material ordering, operation plan adjustment, and accounting management; wherein the step of grouping the apartment complexes according to preset criteria comprises: creating a first group for maintenance efficiency based on facility aging, facility type, and number of households for each apartment complex; creating a second group for management cost reduction based on energy usage and maintenance cost patterns for each apartment complex; creating a third group for service operation efficiency based on resident age groups, income levels, and lifestyle patterns for each apartment complex; and setting data processing weights corresponding to management objectives for each group based on the first group, the second group, and the third group.The step of generating equipment condition prediction information, which includes the probability of anomalies occurring in each of the equipment and the timing of maintenance requirements, comprises: extracting change patterns of current, temperature, pressure, and vibration values ​​from the equipment condition data; inputting the extracted change patterns of the equipment condition data into the artificial intelligence model and calculating an anomaly indication score according to preset criteria by comparing it with past normal state patterns; calculating the probability of anomalies occurring according to preset criteria by combining the calculated anomaly indication score with the equipment-specific aging index, usage frequency, and recent inspection cycle; predicting the timing of maintenance requirements and the scope of maintenance work by applying preset equipment importance weights to the calculated probability of anomalies; and generating equipment condition prediction information including the probability of anomalies, the timing of maintenance requirements, and the scope of maintenance work, and reflecting the generated equipment condition prediction information in grouped complex-specific management data. The step of adjusting the operation plan for each grouped complex comprises: calculating the risk level of each equipment based on the predicted probability of anomalies and the timing of maintenance requirements of the equipment; setting priorities for security, cleaning, and common area management tasks based on the calculated risk level; and, based on the set priorities, according to preset criteria The method includes the steps of: adjusting the deployment of security personnel, cleaning personnel, and inspection personnel, and automatically generating a work schedule by considering the estimated time required for each task, the number of available personnel, and movement paths; calculating an efficiency index for each task according to preset criteria to evaluate the management efficiency of the generated work schedule; regenerating the work schedule if the efficiency index is below a preset threshold value; and adjusting the work schedule where the efficiency index is above a preset threshold value by reflecting it in the grouped complex operation plan, wherein the step of automatically generating accounting management information for each grouped complex isThe method includes the steps of: confirming the timing of maintenance required and the scope of maintenance work for each facility based on the facility status prediction information, and determining the time required for the work by querying a preset work time value corresponding to the work scope; confirming the work schedule for each personnel based on the complex operation plan, and calculating labor costs based on the time required for the work and a preset labor cost unit price; calculating maintenance costs based on the predicted maintenance work scope and a preset unit price for each task; calculating material costs by confirming a preset material unit price based on the generated material order information; calculating a reserve fund by applying the reserve ratio to the sum of the maintenance costs, labor costs, and material costs based on a preset reserve ratio; calculating the total amount of accounting items by summing the maintenance costs, labor costs, material costs, and reserve fund; and determining whether the calculated total amount of accounting items exceeds a preset budget limit by comparing it with the calculated total amount of accounting items, generating budget alert data if it exceeds the limit, and reflecting the total amount of accounting items in grouped complex-specific management data if it does not exceed the limit; and the step of calculating a risk index for each zone within the multi-unit housing complex to set intensive patrol zones and generate a patrol plan. and further comprising the step of reflecting the generated patrol plan in the grouped operation plan for each complex; and the step of setting the intensive patrol zone comprises: a step of calculating a potential risk index for each zone by verifying past alarm history, crime frequency, and facility failure history for each pre-set zone within the apartment complex based on the apartment complex database; a step of collecting real-time video data from CCTVs installed in the apartment complex to extract real-time brightness data for each installation zone of lighting equipment including streetlights and security lights; a step of analyzing the video data collected from the CCTVs to detect abnormal events according to pre-set criteria and generating real-time alarm data; and the potential risk index,The method comprises the steps of: integrating real-time brightness data and real-time alarm data for each zone to calculate an integrated patrol risk index for each zone; setting zones where the integrated patrol risk index is above a preset threshold value as intensive patrol zones; and generating a grouped patrol plan for each complex by calculating the number of patrols, patrol routes, and patrol times according to preset criteria for the intensive patrol zones; and further comprises the step of predicting the lifecycle of facilities for each complex grouped based on the facility status prediction information and generating replacement and maintenance schedules; wherein the step of predicting the lifecycle of facilities and generating replacement and maintenance schedules comprises the steps of: calculating the cumulative operating time and average load rate for each facility based on operating time data, load rate data, and environmental variable data collected through IoT sensors installed in each facility within the multi-unit housing complex; comparing the cumulative operating time, average load rate, and environmental variable data with preset standard life values ​​for each facility to calculate the remaining lifespan for each facility according to preset criteria; comparing the results of the remaining lifespan calculation with past failure history data and manufacturer warranty period information to generate early warning data for facilities with a high probability of early failure; and calculating the failure probability based on failure history data of similar facility types. A step of calculating and determining the equipment replacement priority using the failure probability and remaining life data; a step of determining the equipment replacement timing and maintenance cycle according to preset criteria based on the replacement priority and remaining life of each equipment; a step of collecting performance data through the IoT sensor and calculating a performance deviation index by comparing it with a reference performance value; a step of automatically recommending a response manual by referring to preset failure cause data by equipment type when the performance deviation index exceeds a preset reference value; and a step of calculating a risk grade for each equipment by integrating the remaining life, replacement priority, maintenance cycle, performance deviation index, and response manual information for each equipment.The method further includes the step of predicting the life cycle of the equipment based on the above-mentioned risk grade and maintenance cycle, and generating a replacement and maintenance schedule according to the predicted life cycle and risk grade; and the step of analyzing operational efficiency between grouped complexes based on the above-mentioned management data for each grouped complex, and generating a performance report based on the analysis results; wherein the step of generating a performance report based on the above-mentioned management data for each grouped complex comprises: a step of extracting maintenance cost, labor cost, and material cost data from accounting management information for each grouped complex, and calculating a cost reduction rate for each complex by comparing it with the average value for each complex within the group; a step of extracting work schedule and personnel allocation data from operational planning information for each grouped complex, and calculating a manpower input efficiency index according to a preset standard; a step of verifying material unit price and delivery date data from material ordering information and material management database for each grouped complex, and calculating a material procurement efficiency index for each complex by comparing it with the average value within the group; a step of calculating a comprehensive operational efficiency index for the grouped complex by combining the cost reduction rate, manpower input efficiency index, and material procurement efficiency index; and a step of determining whether efficiency has improved by comparing the comprehensive operational efficiency index with a preset standard value, and if it falls below the standard value, providing data requiring improvement. A method for predicting the status of multi-unit housing facilities based on artificial intelligence and IoT sensors and for integrated management of accounting and material-linked operations by complex, comprising the steps of: generating a performance report including the comprehensive operational efficiency index, cost reduction rate, manpower input efficiency index, and material procurement efficiency index; and transmitting the generated performance report to a terminal of a multi-unit housing manager. Claim 2 delete Claim 3 delete

Images