AU2021105481A4 - Construction operation, monitoring, maintenance planning and future risk prediction of bigger constructions using Artificial Intelligence based Internet of things - Google Patents

Abstract

Construction operation, monitoring, maintenance planning and future risk prediction of bigger constructions using Artificial Intelligence based Internet of Things Abstract: In building maintenance, facilities managers generally use reactive or preventative maintenance techniques. Moreover, there are certain drawbacks, such as reactive maintenance's inability to avoid breakdown and prevention maintenance's inability to forecast the future state of MEP components as well as fix them in advance to improve facility life. As a result, the goal of this research is to use sophisticated technologies and a predictive maintenance plan to address these constraints. The use of the Internet of Things (IoT) as well as building information modeling (BIM) can increase facility maintenance management (FMM) efficacy. Despite major attempts to integrate BIM as well as IoT to the architectural, architecture, constructions, and facilities management (AEC/FM) industries, IoT and BIM connectivity for FMM is even in its infancy. A data-driven preventive maintenance scheduling architecture dependent on IoT and BIM technologies for FMM, comprising an application layer as well as an information layer is created to give an improved management plan for construction activities. The information layer collects as well as integrates data from BIM configurations, FM systems, and IoT networks, whereas the application layer includes 4 parameters for preventive analytics: (1) monitoring as well as fault alarming method, (2) evaluation method, (3) situation forecasting method, and (4) planned maintenance method. The SVM and ANN machine learning techniques are utilized to forecast the future state of MEP elements. Thus, the machine learning algorithms indicate that continually modified data of information layers, along with the application layer will accurately forecast the future state of mechanical, electrical, plumbing (MEP) elements for maintenance planning. 1

Description

Construction operation, monitoring, maintenance planning and future risk prediction of bigger constructions using Artificial Intelligence based Internet of Things

Description

Field of the Invention:

This invention is intended in developing a predictive model to identify the future risk in the bigger constructions. This invention uses the Internet of Things (IoT) and Artificial Intelligence (AI) technology namely machine learning (ML) algorithm. Thus, this invention helps to improve the intelligence and life span of the building by providing better constructional operation, continuous monitor and maintenance of the buildings.

Background of the Invention:

Approximately 75% of the annual FM expenditures are spent on building upkeep. To be able to save money on upkeep, you should use the proper maintenance methods, and in turn, potentially increase the life of your building's materials.In the current state of building maintenance, reactionary and preventative maintenance are used. When FM staff does reactive maintenance, they do so after a failure has occurred. Preventative maintenance is a schedule strategy that allows facility management personnel to examine or upgrade building components at preset intervals or times. Reactive maintenance is unable to avoid failure, whereas preventive maintenance is unable to foresee future conditions and fix components ahead, therefore extending the life of construction materials. Predictive maintenance, also referred to as condition-based monitoring, seeks to predict impending failures as well as eventual deterioration depending on patterns in component states discovered from historical data, allowing for early intervention. While building components will still be in excellent working condition, predictive maintenance may solve the foregoing drawbacks by anticipating potential breakdowns and fixing components in advance to increase the service life. This method is heavily reliant on sensor-collected and communicated operational data. Examination (including periodical surveys), as well as continual surveillance (including sensor tracking), are the two major methods for gathering data information in a facility. Preventive maintenance selection necessitates the combination of a variety of data sources, including data collection, regular maintenance, production schedules, causation, and the knock-on impact of failures, among others. Solutions using computer-based equipment are used to make facility maintenance management (FMM) activities more effective.

A huge wastewater treatment facility with a well-integrated BIM model provides intelligent construction, operation, and maintenance. BIM is a technology that can help software developers, architects, and building managers understand, manage, and display data at all phases of building development. As a result, BIM enables engineers to accurately comprehend and react to various forms of architectural data while also serving as a basis for collaboration among the structure and construction teams. As a result, BIM performs a critical role in boosting productivity, lowering costs, and speeding up the construction process. They provide a case study in which BIM was utilized to intelligently design a massive wastewater-treatment plant (WTP). Intelligence modeling and prototyping methodologies, construction maintenance and modeling, digitized distribution systems. BIM-based Internet of things operation and maintenance, as well as environmental monitoring, is part of the procedure. The WTP's construction, distribution, and operations, and maintenance procedures all are digitalized using BIM.

A new artificial intelligence (AI) intelligence layer for buildings for more energy efficient operations. A new era of municipal energy efficiency may result from increased attention to smart buildings and incorporating sensors, big data (BD), and artificial intelligence (AI). Al can help manage, ensure, and automate energy use in smarter buildings. as well as building management systems (BMS).In addition to discussing the precepts and objectives of A-dependent modeling approaches commonly utilized in building energy usage identification, evaluation architecture is presented and utilized to evaluate the latest research on this topic and around the significant Al domains, such as energy, convenience, structure, and maintenance. Lastly, the article offers a review of outstanding issues and upcoming research areas in All's use in smart buildings.

A hybrid artificial intelligence model predicts the likelihood of risk delay in building projects. Project delays are one of the most significant construction industry problems. The solutions to complicated, unexpected issues presented by artificial intelligence (AI) frameworks have already been proven. to identify the potential for delay problems, a group of researchers has developed a technique that integrates Random Forest classifier optimization with Genetic Algorithm optimization (RF GA).The origins, as well as factors of delay issues, are initially recognized. To assess the influence of delay causes on performance measurement, a set of questions is used. The hybrid model that has been built is trained to utilize data from past construction works. A statistical measurement of performance indices is used to compare the suggested RF-GA to the traditional variant of an RF model. The created hybrid RF-GA method's attained results indicated an excellent resulting performance regarding accuracy, classification error, as well as kappa.

Objective of the Invention:

1. This invention helps in improving the constructional quality and intelligence of the buildings. The operational values of the buildings can also be enhanced. 2. The main aim of the invention is to incorporate the buildings along with the artificial intelligence and IoT. The artificial intelligence uses the machine learning algorithm. 3. The proposed algorithm computes the risk based on the previous data and provides an accurate result. Thus, the smart buildings are prevented from the dangerous hazards.

Summary of the Invention:

Urban systems are made up of a variety of multiple-stage dynamic entities which play an important role in society's structure. These entities are responsible for the regular functioning of society, numerous features of urban morphology, and the management of social constructs. The technical structures which are accountable for the seamless functioning of the interrelated components in the urban environment must be flexible, efficient, and long-lasting. The creation of smart cities including huge residential and business structures is projected to rise as the world population expands and urbanization accelerates. Smart cities rely heavily on information communication technology, as well as Internet of Things (IoT) methods. The adoption of such technologies in current cities is accelerating the shift from the 20th-century city idea to upcoming smart cities. Yet there are substitutes to several features of conventional city life, the entire superstructure of contemporary life is based on it.

In modem times, significant advancements in computerized communication technology have contributed to the application of artificial intelligence (AI) in urban building transformations. Design optimization in a diversity of urban contexts has been automated owing to smart control provided through the use of computer models. The broad range of potential of Al in urban design, as well as its capacity to operate autonomously, is driving its fast growth. Although many people still think of Al as a farfetched concept that fits in science fiction, A-based technologies are becoming an actuality in modem infrastructure. Recent advances in Al have aided in the resolution of urban issues such as impulsivity, ambiguity, mismatch, complication, and evolutionary inconsistencies. Artificial intelligence (AI) has provided us with automation systems for designing, analyzing, simulating, controlling, diagnosing, and supplying safe buildings in metropolitan areas. These technologies can allocate constructing resources and manage supply and need to keep the grid running efficiently and cost-effectively.

Al is unquestionably well equipped to flexibly control urban building systems based on contemporaneous data. It is demonstrated to be extremely competent in operating the autonomously grid system management by effectively optimizing utilization via improved optimal supply and request management as well as performing energy generation auto-maintenance. Artificial Intelligence can also help to decrease peak demands and keep the grid stable. Al-based identification method can improve dispatch and reduce power production operating costs by using energy need and supply predictions, meanwhile, power plants satisfy need and run inside the energy network's operational limitations over lengthy periods of fluctuating circumstances. Demand responsiveness with end-use energy consumption is enhanced, and energy costs on the facilities are lowered, due to the application of artificial intelligence (AI) as well as machine learning (ML).

Al as well as IoT is currently at the forefront of all major technique companies' agendas. The advent of construction technology research demonstrates that necessities are changing throughout time. In the coming years, the necessity for protection will become even more important. The basic criteria for system functioning will be adaptability and dynamics. The highest level of efficiency would be prioritized. One of the most widespread issues globally in the construction business is to make buildings more efficient to benefit the occupant's health, security, and convenience. Al appears likely to become the unique option that can achieve the highest synergy among construction performance, comfort, flexibility, protection, and security while also being long-term sustainable. Buildings become living creatures networked, smart, and adaptive to the fluctuating demands of their providers and customers because of artificial intelligence. The interaction among internal customers and producers is managed by a smart constructional management system. By improving the pace of energy savings, enabling on-site generation, identifying, and eliminating operating problems, as well as regulating and assuring continuous energy savings, Artificial Intelligence (AI) can assist solve the obstacles of constructing efficient buildings. Combining sophisticated data, onboard graphical diagnostics, as well as cloud connection will ensure the advantages that come with seeking maximum energy savings. Having various control systems in a structure causes complexity and, therefore, results in high expenses. As a result, while installing Al technologies, employing a digital transformation mentality will save operating expenses whereas increasing the building's whole performance.

support vector machines (SVM), Artificial neural networks (ANN), as well as Markov chains, are some of the machine learning methods that may be used to forecast the state of building elements. Because of their capability to anticipate nonlinear time sequence trends, ANNs having been employed as a decision support system. The capacity of artificial neural networks (ANNs) to record and maintain nonlinear dropout sequences has been widely explored and reported. Identification of nonlinear time sequence is employed. ANNs are being proven to outperform existing conventional auto-regressive models. In their capacity to effectively understand nonlinear characteristics of a time sequence. ANNs vary from standard statistical approaches, and they are frequently utilized in prediction. Models are created, for instance, to assess and forecast the state of pipelines depending on a variety of parameters, especially corrosion. To assess the stability and life span of a facade coating, researchers employed multiple linear regression assessment and artificial neural networks (ANN), as well as statistical models to explain facade deterioration.

SVM is a commonly utilized statistical learning theory-based classification algorithm. The effectiveness of SVMs and ANNs for evaluating the fundamental state of sewers is assessed, and it has been discovered that SVMs and ANNs distinct benefits in evaluating the structural behavior. Using unique cases in training and testing samples, SVM has a higher likelihood of a parameter setting influencing the results. The Markov chain method has been utilized to forecast the future state of bridge elements and to estimate its service life. The Markov chain method has two drawbacks: (1) Discrete parameters were utilized in the model. (2) It presupposes that the future state is solely determined by the current state, not by the state of the past. Because of these drawbacks, the Markov chain is ineffective for various construction elements, including the HVAC system.

However, its execution is frequently hampered by technical constraints arising from the solution's complexities, and legal and budgetary constraints. The SVM and ANN algorithms are chosen as machine learning techniques to forecast the upcoming state in this work depending on the collected datasets. Furthermore, the effectiveness of both SVM and ANN is dependent on the number of databases utilized to train the model networks, rises with the number of databases. The amount of information generated has a direct relationship with the model's quality. As a result of our research, the prediction procedure would be improved, and prediction models become data-driven approaches depending on the contemporaneous data.

Detailed Description of the Invention:

The arrival of smart buildings is the highest potential use of Al in city energy sources. Smart buildings are generally outfitted with a variety of actuators, sensors, subsystems, as well as a variety of modem and smart autonomous monitoring and controlling tools that may be used to save energy. These energy consumption decreases are complemented by decreases in greenhouse emissions. This is demonstrated how smart buildings can safeguard the environment, reduce building operating costs, and conserve energy in metropolitan settings. The smart building is a construct that exemplifies today's infrastructure. It has autonomous controlling systems and makes use of data to improve the building's performance as well as the degree of convenience for its occupants. It offers a cyber-physical system (CPS) that is responsible for connecting the cyber world along with the virtual world comprises different electric appliances, electronic gadgets, controllers, sensing devices, and metering elements.

The smart building idea is founded on the accurate notion which enables free relationships among the two worlds to provide benefits. The influence of Al techniques in smart buildings is becoming increasingly evident as technology develops. The advent of progressively compact, more energy-efficient sensing as well as communication protocols, along with simultaneous developments in software as well as hardware technologies, has made assessing, monitoring, and engaging with the surroundings easier. This idea, termed the building management system (BMS), suggests that Al can improve not just the functionality of buildings and also their life experience in the coming.

Despite building automation as well as energy management practices are mostly used for monitoring as well as alarming. With the growing confluence of smart buildings, creating a single analytics platform to deliver greater insight from the aggregated data has never been more important. Building energy usage is monitored, collected, controlled, evaluated, and managed using artificial intelligence. It monitors and manages energy use, reducing this during peak times, identifying and signaling issues, and detecting equipment breakdowns before they happen. ML algorithmic approach and blockchain are used to allow active client involvement in requirement response programs utilizing Al-based methods. Although heating, ventilation, and air conditioning (HVAC) methods keep people comfortable inside, they also lead to increased energy usage. District heating powered by Al adds unrivaled flexibility to the next energy system in bigger construction.

A common heat exchanger loops installed in the ground of constructed buildings transmit may link the heat pumps of many smart buildings effectively and knowledgeably. Since electrification raises the controllable count components as the energy system grows highly complicated, Al can help automate district heating. This can forecast client heat demand, direct storage consumption, and assist the controlling room in making the best use of resources. The predictive maintenance architecture incorporates modem techniques such as IoT, BIM, and a facility management system. The information layer as well as the application layer is the two levels of the architecture. The application layer comprises four stages for (1) monitoring as well as faults alarming, (2) evaluation, (3) identification, and (4) planned maintenance, whereas the information layer connects diverse types of data from BIM methods, IoT sensor network, as well as an FM system.

Data collection:

The needed data is separated into 3 categories: (1) Construction facility geometrical and semantics data (non-graphical data), including kind, materials, dimensions, capacity, location, and implementation era. These are statistics that are extracted from the BIM method and show the inherent characteristics as well as the intrinsic degradation trend over time. (2) Information on facility conditions, comprising sensor data of pressure, temperature, and rate flow received via an IoT sensor networking, as well as operational data of every essential component. Sensors have been utilized to track the functional status of essential elements as well as the development of sensor information that can reveal the frequency of unusual occurrences and element consumption patterns. (3) Maintenance-related information and documentation obtained from FM networks, such as inspection reports and previous maintenance reports. Maintenance-related papers include characteristics such as utilization age, annual inspections, abnormal inspections, minor maintenance instances per year, as well as major maintenance instances per year. While functioning as the input for the predictive model, these specific characteristics are crucial for analyzing the current state and forecasting the upcoming condition.

A repository of loT sensor network data

A sensor networking infrastructure is constructed while the operational phase is ongoing to gather sensor information from building services and the surrounding environment. Information is collected from the DDC network and then the status and outside external variables are determined by decoding the signal. Furthermore, a dock-equipped plug-in for the BIM technique is being developed that makes it possible to see and store sensor data in the situation databases while also using the BIM method.

The development of data integration and visualization

In facilities management, there are three different kinds of data that are integrated and presented in 3D BIM models. These models are then used for facility evaluation and efficient optimization. COBie-based FM and BIM data fusion, IFC-based sensor object with IfcSensorType extensions, sensor data format development, and sensor modelling and data analysis comprise the data aggregation and presentation process.

Highly accurate monitoring and signaling modules

Predictive maintenance consists of the gathering and analysis of key element characteristics to detect both normal element changes as well as machine health trends. Preventive analytics begins with condition monitoring and then moves on to fault alerts. For starters, the developed plug-in makes it easy for FM technicians to get sensor information, to inspect each piece of equipment, and to be engaged in automatic control data patterns monitoring. Fault alerting is often associated with aberrant MEP element occurrences, such as unexpected vibrations in a machine or when the chiller's temperature deviates from normal. Preventive analytics theory begins with the assumption that when a parameter in an MEP element reaches a certain threshold, an alert or notice is sent to let you know that something is wrong with the element. The facilities management handbook, on-site checks, and reference to statistical data trends are tools that facilities managers should use to evaluate if an aspect is problematic.

Condition prediction module:

The goal of prediction management is to aid in the planning of maintenance by identifying faults and forecasting the state of building materials. Conditions prediction would be based on conditional monitoring information acquired through its sensors; FM information gathered via the FM system, information of BIM, as well as the condition, is evaluated. To forecast the future state of MEP elements, SVM and ANN algorithmic approaches are chosen as machine learning methods. 15 factors are acquired from several systems, including FM structures, BIM models, and IoT sensor networking, as input towards this prediction approach. Furthermore, each MEP element has its own set of variables. Pressure, temperature, and flow rate sensors, for instance, are used to check a chiller, whereas a vibration sensor is used to manage an elevator. The conditioning index of MEP elements in buildings, as well as triggers and alerts for needed maintenance activities, are among the outputs of this prediction step.

Construction operation, monitoring, maintenance planning and future risk prediction of bigger constructions using Artificial Intelligence based Internet of Things

Claims:

This invention involves the development of algorithm for the prediction of the risk in the smart constructions and it involves the following claims:

1. This invention involves the integration of Internet of thing (IoT) and artificial intelligence (AI) specifically the machine learning (ML) algorithm. i. From claim 1, The Artificial Intelligence has the capacity to tackle the problems that are complex and unpredictable in nature. ii. From claim 1, The ML algorithm reduces the energy cost and also enhance the responsiveness of the user energy. 2. This invention involves the development of the hybrid AI-IoT risk prediction model for investigating against the time delay. This hybrid model uses the trained and tested data from the past construction work as the input. i. From claim 2, Al appears likely to become the unique option that can achieve the highest synergy among construction performance, comfort, flexibility, protection, and security while also being long-term sustainable. 3. The integration of the IoT and Al technologies involves the following process. i. From claim 3, Data is collected from the IoT sensors and other data that are needed are geometrical and semantics data of the construction, data on facility conditions, and maintenance-related information.

From claim 3, the collected data is integrated, and then visualized. Finally, the trained and tested data is checked for the condition monitoring and false alarm.