Innovation Challenges

Traditionally, progress monitoring of construction projects is often influenced by human factors, as project steps – inspection, recording and interpretation – are done manually by different stakeholders. The resulting reports are often not reliable and require further scrutiny by supervisors and higher-level managers.  

Reliable real-time progress monitoring is important to identify issues and potential delays in a timely manner, so that responses like manpower redeployment can be planned and executed to resolve them.

Surveillance tools such as CCTV are being adopted more widely on the construction sites, but these tools have limited usefulness due to the reliance of humans to carry out the remote monitoring.

We are interested in AI-based video analytics solutions that can use the footage from existing or new CCTVs to produce useful insights on on-site productivity. Instead of having site personnels to perform frequent checks on the site, the solution would intelligently monitor the site and alert users of any issues or anomalies.     

Stage completion and floor cycle are the priority areas for progress and productivity tracking.  The tracking of metrics related to manpower, machinery, logistics and material handling are also of interest. 

The solution should fulfil the following specifications in order to be considered successful:

• It must integrate with the existing CCTV system. The majority of Tiong Seng’s current projects utilise video surveillance systems from Dahua Technology.
• It must be able to accurately identify and track the following Reinforced Concrete work activities to quantify the stage completion of a floor as the percentage of completion: 1) Reinforcements works, 2)Formwork installation, 3) Concreting works and 4) Formworks removal
• It must be able to compare the actual progress against projected progress based on a seven or ten-day floor cycle time, and identify issues and potential delays. 
• It must include methods to boost connectivity to allow for real-time monitoring and analytics.  
• The accuracy of the models for the various use cases should be at least 85%.
• It should be able to analyse the duration taken for different stages completed within a floor.

The following are some additional requirements that could further enhance the proposed solution

• The solution could include APIs for interoperability with other enterprise tools and platforms.
• It could be compatible with Building Information Modelling (BIM) to allow for BIM models to be overlaid on real-time CCTV footage or captured images.
• It could involve the installation of new surveillance tools if it can boost the accuracy of the models in a cost-effective manner.  

The solution must include a digital platform that analyses the data captured, visualises the data on a dashboard, produces daily progress reports, and alerts key stakeholders on issues and delays. The platform should integrate the existing live CCTV streams through a RTSP setup.

Expected Outcomes

An AI-based video analytics solution can monitor and track structural work as well as formwork progress from the plan view and elevation view of a construction site. The solution allows for early identification of delays and risks that could cause budget overruns

Learn More

Buildings and construction accounts for 39% of global greenhouse gas emissions. Firms in the construction industry, including Obayashi, are increasingly committed to work towards carbon neutrality. As part of Obayashi Sustainability Vision 2050, measures introduced to reduce carbon emissions include:

• Reducing the fuel used by promoting new technologies and labour-saving construction; and
• Developing and practically using energy-saving construction methods and low-fuel consumption and electric construction machinery.

In order to determine the effectiveness of such measures, it is important for Obayashi to first track the current carbon emissions of their sites. Currently, there are no established practices of collecting site-level data related to carbon emissions. If Obayashi were to implement new practices that involve manual data collection, it would be too time-consuming and also produce data that is unreliable due to human factors.

Obayashi wants to reduce both direct and indirect carbon emissions. Indirect carbon emissions are contributed by other companies operating on their sites, such as subcontractors and suppliers

We are interested in solutions that can automatically collect site-level data related to carbon emissions. The data of interest include:

• Carbon emissions produced by logistics vehicles, for example, dump trucks and lorries. This can be estimated by determining the number of vehicles operating onsite and their fuel consumption. 
• Carbon emissions produced by construction machines, for example, tower cranes and excavators. This can be estimated by determining the number of machines operating onsite, the type of operation, and their energy consumption.

The solution can adopt AI-based visual analytics methods on existing site surveillance systems to collect the data mentioned above for all companies involved without requiring site-wide modifications of vehicles and machines. We are open to adopting new surveillance systems that can support effective and accurate collection of data. 

In order to better plan measures and promote collective action by all stakeholders, it is useful if the data collected can generate the following insights:

• Breakdown of the carbon emissions based on company
• Breakdown of the carbon emissions based vehicle or machinery type and model
• Identification of operation abnormalities that could lead to higher emissions, for example, smoke generated from a faulty machine. 

The solution can be integrated with project management or logistics management platforms to produce more accurate carbon emissions data on the project sites. It would include a dashboard to visualise the data collected and produce useful breakdown of carbon emissions.

Expected Outcomes

Concrete grinding is done to remove the defects on the concrete surface, such as unevenness, bulges and drips, so the quality of the architectural finishes is not affected. Currently, concrete grinding is performed manually by workers using handheld grinders. This operation is not only labour intensive but also poses health and safety hazards to the workers, especially during long periods of grinding work. In order to grind concrete surfaces that are beyond the reach of the worker, for example, walls with heights between 1.5m and 6m, and ceilings, work-at-height equipment is needed. Working at height adds complexity, requires more manpower, and also presents additional safety risks that need to be managed.

Often, the surface finish of grinded concrete surfaces still requires additional treatment, such as skim coating and plastering, before it is ready for a paint finish. As a guide, Concrete Surface Profiles (CSP) were developed by the International Concrete Repair Institute (ICRI) to divide concrete surface treatments into 10 classifications. They are named from CSP 1 to 10 – with CSP 1 being the finest and smoothest surface, and CSP 10 being the roughest and most uneven surface.

For concrete grinding, CSP 1 or CSP 2 is the expected level. Also, the concrete surface should be ready to receive finishes, for example,  paint finish, of not more than 10mm.

We are interested in robotic solutions that are able to partially automate concrete grinding works as a start and can later be developed into a fully-automated solution. In the early development stages of the solution, the workers can deploy it to do the grinding for a chosen section of the wall or ceiling, and after the work for that section is complete, return to turn on the machine for the next section. This then frees up the workers to perform other activities while eliminating the safety risk associated with grinding. 

The current intent is to use the solution to grind the surfaces of a cast in-situ or precast wall that are casted from a typical concrete formwork. The surface profiles of such concrete surfaces are usually even, except for some minor bumps and drips. 

The solution should fulfil the following requirements in order for it to considered successful:

• It must accurately grind all of the assigned surfaces to produce a consistent surface quality.
• It must accurately identify the quality and defects of the concrete surface and adapt the grinding accordingly. 
• It must be able to perform concrete grinding work on a surface area of no less than 8m² an hour. 
• It must grind surfaces of up to 3m in working height 
• It must allow for Interchangeable parts, such as different types of grinding discs and grinding activities. 
• The resulting quality of work from the solution must comply with International Concrete Repair Institute’s guide on CSP, where the quality standards for internal and external finishes works shall adhere to CSP 1 to 2, and also to comply with BCA CONQUAS 2019 (Appendix 1, Architectural Finishes, Item 2a - Internal Walls & Appendix 2, External Wall, Item 1 – General Requirements ) as further supplementation.
• The grinding must be performed in full consideration for safety on site and environmental impacts.
• The solution must support the capture of data related to the work done, such as location, duration and errors, for progress tracking and inspection purposes.  
• It must be able to operate on site during the construction phase where there are only construction power sources i.e., on-site power generators.  
• The solution must be able to fit into material hoists or other equivalent vertical transportation, or work-at-height equipment, with consideration for its size, weight, and portability.

It is also desirable for the solution to meet the following performance specifications:

• It is designed to access heights of up to 6m or to integrate with the common work-at-height equipment for elevation.   
• It should ideally improve the quality of works to reduce or eliminate surface treatment works required in order to prepare for the paint finish.
• It should be able to traverse and park on uneven, sloped or stepped areas. 
• It should consider how grinding can be done for the following joint areas : 1) joint areas between vertical or horizontal elements (e.g.: slab soffits, wall and floor joints and 2) joint areas between vertical-vertical elements or horizontal-horizontal elements (e.g.: areas between walls and slabs)

The ideal solution can autonomously perform concrete grinding, change grinding parts, and navigate a building floor. For a start, the solution should complete the concrete grinding work on a surface area without human intervention except for deployment, redeployment and termination, and do so accurately and quickly (over 8m² an hour).

This reduces or eliminates the need for manpower, allows for remote execution and reduces the total cost and time needed to prepare a concrete surface for receiving architectural finish.

Grouting is the process of filling the gaps in between tiles (ceramic, marble, granite, etc.) for the final finishes of floor and wall. It is a labour-intensive and repetitive process that demands time and skill to achieve the desired finish quality. 

Here are the steps involved for the current practice of grouting:

1. Mixing of grout material as per the manufacturer recommendation.
2. Preparation and cleaning of the area.
3. Removal of dirt and unwanted substances from the gap between two tiles. 
4. Application of grout into the gaps between tiles: the grout is packed into the gap using a plastic card or rubber pad. 
5. Removal of excess grout and cleaning of tile surface.

We are interested in robotics solutions that can automate the entire grouting process for floor tiles in an integrated manner. For a start, prioritise the automation of the steps in the grouting process that have the most potential to generate productivity gain and improvements to quality. Step 4, as described above, is of top priority for automation, followed by Step 5 and Step 1. 

The solution for Step 4 should fulfil the following specifications in order to be considered successful:

• It must accurately identify the tile gaps and autonomously navigate a room of up to 20m² to apply grout for at least 50% of the room. 
• It must consistently apply grout for different tile arrangements. The tile sizes used in the arrangements could be as small as 150mm in width. 
• It must also consistently apply grout to fill the tile gaps of 1 to 2.5mm with different depths (depending on the tile thickness). 
• The grout is considered to have been applied consistently when there are no gaps or excess in the application. 
• The solution must be able to achieve a recessed joint profile. For a flush joint profile, the robot should achieve as close to what is practically possible.
• It should be able to tailor to different types and composition of grout mixtures and adjust the grout application process accordingly.
• It should be able to apply grout to the corners and sides of the floor tiles that interface with the wall

The robotics solution could take innovative approaches to simultaneously conduct the different steps in the grouting process, instead of a step-by-step procedure taken by human workers. This should reduce the time taken to complete the grouting process and even reduce material wastage (pre-mixed grout is discarded if it hardens). The following solution features are desirable:

• The solution removes excess grout from the gaps and tile surfaces before the grout hardens.
• It is able to carry the different mixture components of grout (for example, cement, water and latex) and mix the grout using a just-in-time approach.

The solution must include a simple digital interface for workers to easily program the robot to perform the grouting works for different projects and monitor the work progress. 

Expected Outcomes

Quality inspections are required at different stages of the precast fabrication process to ensure that the precast components are constructed accurately according to the required specifications. The inspections are done before the concrete is poured (“pre-pour”) to check the mould, reinforcements, lifting hooks, splice bars, and connectors; and also, after the concrete has hardened (“post-pour”) to check the resulting concrete size, shape and quality, and other items on the checklist (see below).

These inspections are done by Resident Technical Officers or Resident Engineers representing the fabrication plant or consultant. The fabrication process needs to pause for inspection before it can proceed to the next stage. The inspectors need to be present on site to visually inspect the precast components and verify the measurements by taking those measurements with measuring tape and the assistance of other workers. They also need to capture photographic evidence for documentation purposes.

Greyform has already adopted robotics to automate the fabrication of various 2D precast components such as wall, column and slab, and is now looking for solutions to automate the quality inspection process.

We are keen to explore the use of advanced imaging and laser-based technologies together with AI to accurately inspect, measure and analyze the precast components to generate the information needed for inspection. The solution must be able to:

• Measure dimensions with a precision of +2mm for panel dimension and +10mm for opening/recess.
• Compare the actual precast components against their Building Information Modelling (BIM) to identify any deviations or defects during post-pour. The comparison is done with the consideration of allowed tolerance for each inspected item.
• Differentiate the rebar mesh from the bottom slab or metal pallet, taking into consideration that the different items can have a similar colour.

The solution should also consider how it can be integrated with Greyform’s Automated Pallet Circulation Plant system.

Expected Outcomes

An AI-based automated inspection system utilizes imaging and/or laser technologies to determine whether precast components are fabricated accurately. For each inspected item, the system produces an inspection report supported by photographic evidence with the measurement (if applicable) and highlights any abnormalities for further action.

The solution reduces the manpower required to support the inspectors by eliminating the need for manual measurement and capturing the inspection data during the fabrication process, so that the process does not need to stop for inspection. In the future, the solution could support remote inspection procedures.

Learn More

Chuan Lim Engineering

Noise barriers are installed along expressways, roads, MRT tracks and construction sites in Singapore to dampen the noise from vehicles, trains and machinery. In order to enable quieter living environments, there are plans to install more noise barriers, for example along the North-South Corridor and along MRT tracks in places like Joo Koon, Bishan and Paya Lebar. 

Workers use a work-at-height equipment to carry and slide the panels into column structures. For a noise barrier that is 4.5m high, this process is repeated till each column is filled with nine panels. The current method requires two men on the boom lift to perform the installation and at least two men on the ground to pass the panels and operate the work-at-height equipment.

What We Are Looking For

We are interested in robotics solutions that can automate the installation of noise barrier panels. The solution should help reduce the manpower and machinery involved in the panel installation process.

The following are requirements and considerations that are important when designing the solution:

• The solution must be able to transport the panels to the point of installation in a manner that does not cause damage to the panels or distort their shape. Each panel has a length of 2m and a height of 0.5m.  Its thickness varies between 8mm to 60mm, and depending on type, it weighs up to 12kg. 
• The solution must consider how the panels are collected for installation. It could include a panel storage system to ease the collection process.
• It must accurately identify the profile of the column structure and slide the panel to the correct height, depending on which panels have already been installed. 
• The panel installation process should take less than 10 minutes per column.
• The solution should ideally be able to automatically move onto the next column once the installation for a column is complete. 
• It could be designed to access a height of at least 5m or be integrated with work-at-height equipment, for example, a boom lift. 

The solution should include a simple interface to programme and remotely control the robotics solution. 

Expected Outcomes 

A robotics solution can collect and install noise barrier panels for noise barriers of up to 4.5m in height within 10 minutes. The solution ideally reduces manpower required for this process to just one person.

To construct deep base projects and underground structures, a diaphragm wall or a continuous underground wall is used. This method employs Trench Cutter Machines on the ground to excavate a trench along the deep excavation project's peripheral axis while maintaining mud protection.

The Trench Cutter Machine is not easy to operate and relies heavily on the operators’ skill and experience. The operator uses visual observation and monitors the machine’s data to assess the soil condition to determine how best to operate the machine to carry out the excavation. 

The diaphragm wall construction method is widely used in Singapore. There are over 60 units of Trench Cutter machines in Singapore, accounting for 40-50% of all machines in the world. However, there is a shortage of experienced machine operators. Poor operations of the machine impact productivity and the quality of the excavation work, and even cause accidents that might delay the construction timeline. 

For monitoring purposes, the operators also collect data on the excavation operation and the machine from the control panel, and manually input the data into LT Sambo’s platform.

What We Are Looking For

We are interested in solutions that can be retrofitted to the Trench Cutter Machine and can serve to collect and analyse data on excavation operations. The data collected can benefit the construction design process, machine configuration, cost estimation, operator training, and even support future automation of the machine. 

We are keen to collect the following data related to the machine and excavation operations.  The data will then be channelled to LT Sambo’s platform for recording and site monitoring purposes.

• Time (Y/M/D H:M:S)
• Depth (m)
• Speed (cm/min)
• RPM (rpm)
• Direction (˚)
• Pressure (bar)
• Load (tonne)
• Operator behaviour
• Soil characteristics 

Through the processing and analysis of the data, we hope to achieve the following:

• Determine the soil profile of the construction site – Having more accurate records of the soil type and conditions along the depth of the entire trench support more informed decisions regarding construction design, machine configuration, and cost estimation for future projects, especially those located in close proximity.
• Build a knowledge database for training machine operators – Operators can refer to machine operation patterns of skilled operators and learn from them, in order to refine their own performance.
• Lay the foundation for machine automations – The knowledge and data of machine operation patterns can be used to develop machine automations.  

The following are the key challenges that the solution need to address:

• It is not feasible to tap into the machine’s control systems, export the data from machines, or develop APIs in the near future.
• The solution should ideally not tap into the machine’s power systems

When carrying out concrete works in a precast plant, especially in an Integrated Construction and Prefabrication Hub (ICPH) where a Just-In-Time (JIT) manufacturing principle is adopted, it is important to ascertain the early-age strength development of the concrete. This is because components can only be demoulded if and when the minimum strength has been attained. This is usually measured using the conventional concrete cube testing method. Additional concrete cubes are also made to be sent to the relevant authorities for later-age strength measurement.

Concrete cube testing is a manual and laborious method. The test also results in a lot of waste as the crushed concrete cubes from the compressive load test will be discarded.

The results from the concrete cube test are often not an accurate representation of the actual state of the precast components due to the differences in curing conditions. Concrete in small cubes has the tendency to gain strength at a slower rate than the precast components that are larger in size. Thus, there are missed opportunities to demould the precast components earlier and have a faster delivery schedule.

What We Are Looking For

We are interested in concrete sensing technologies that can measure the real-time strength development of concrete with minimal wastage of materials. Strength development of concrete could be inferred by analysing internal temperature changes of concrete to understand the chemical reactions that have taken place.

A new testing approach could reduce or eliminate the need to carry out the conventional concrete cube testing method. It also boosts production efficiency by giving real-time insights into the concrete structure, so that it can be demoulded as soon as the minimum required strength is attained.

The test will be conducted on flat slabs produced in the ICPH. The solution could be based on different sensing approaches, but need to consider the following requirements:

• The sensing technologies must work for precast elements with thickness of 70 to 90mm. Ideally, the technology can work for thicknesses of up to 150mm. 
• Embedded sensors must be small enough to be placed within the rebar structure of the precast components.
• Embedded sensors should ideally operate wirelessly so that there is no additional effort required to find the sensors.
• External sensors must be able to accurately identify the internal conditions, for example, the temperature of the precast element.
• The solution must have a low implementation cost that amounts to <1% of the total component raw cost. 
• The solution should consider the concrete strength development benchmarks according to EN206 and other relevant standards set by local authorities.

The solution must include a tablet-friendly digital platform that integrates and visualises the data from the sensors, and produces insights and alerts that can be easily understood by the on-site workers.

Expected Outcomes

The concrete sensing technology accurately measures the real-time strength development of precast flat slabs, so that the demoulding can be done at the best time to increase production by at least 25% and reduce the risk of defects caused by demoulding before the required concrete strength is attained.

Learn More

Waterproofing involves the application of waterproof membranes or physical barriers to prevent moisture from penetrating building structures. Waterproofing work is a manual process done by skilled workers. The quality of work is highly dependent on the skill and execution of these workers at each step. 

Supervision and timely instructions are important to ensure the quality of work, as the waterproofing project sites could vary in requirements, conditions, and thus complexity. It is time-consuming and costly for the engineers and technical managers to visit the sites, especially those located overseas, to provide supervision and perform quality inspections. Even with site visits, it is still not feasible to monitor and inspect the entire waterproofing process to ensure every step is done with minimal flaws. 

Any flaw in the process can lead to waterproofing failures, which then require costly rectification works or worse, cause damage to the building structure or system. Poor workmanship is the main cause of such failures.

What We Are Looking For

We are interested in solutions that enable the remote monitoring, inspection and/or instruction of the workers. The solution could even improve the waterproofing process to help the workers consistently achieve good quality of workmanship. It would ultimately help to reduce the need for engineers and technical managers to be onsite to ensure the quality of waterproofing work.

The solution could be IoT devices installed on site, attached to the workers or both, and must fulfil the following requirements: 

The solution must include a dashboard that integrates the monitoring information and alerts from different project sites.

Expected Outcomes

An imaging device performs real-time remote monitoring of workers performing waterproofing activities, thereby reducing the need for engineers and technical managers to be onsite. The data collected is analysed, and relevant stakeholders are sent alerts regarding the progress, issues, and abnormalities.

NatSteel

To splice two reinforcing bars (“rebars”), a coupler is used. Conventional couplers work based on a mechanical principle similar to nuts and bolts: the rebars are screwed into a coupler.

Couplers are difficult to install onsite because the rebars need to be precisely aligned with the coupler during the splicing process. The threads on the rebars and couplers are fine, and thus could be easily damaged during the splicing process. 

After the coupler has been installed on the first rebar, the second rebar is rotated and screwed into the coupler (see illustration below). This task is challenging, because the rebar is both heavy and long, and thus the splicing requires at least three men to carry out.

We are interested in alternative designs for the couplers or innovative methods of joining two rebars together.

• The solution must conform to the standards for "Steels for the reinforcement of concrete — Reinforcement couplers for mechanical splices of bar” (SS ISO 15835), especially for the following requirements (more information can be shared upon request to beamp@padang.co) : 1) Tensile strength under static force, 2) Ductility under static force and 3) Slip under static force
• It must eliminate the need to rotate the rebar as part of the installation process, so that less manpower is required. 
• It must be applicable to rebars of different thicknesses, ranging from 16 to 50mm.
• It should ideally not require any threading or other types of processing method to be performed on the rebar. This would save at least 10 minutes per rebar. 
• It should ideally have a quick installation process of less than 10 seconds per joint.
• The introduction of any additional tools would need to be further considered for important aspects, such as space, safety and skills requirements. The solution should ideally not require any additional tools for the installation process.
• A push-and-lock or a press-and-lock system could be considered.

Prefabricated Bathroom Unit (PBU) is a bathroom unit that is pre-assembled off-site, complete with finishes, sanitary wares, concealed pipes, conduits, ceiling, and bathroom cabinets, before it is delivered and installed onsite.

At different steps of the fabrication process for PBU, reference lines and markings are required to assist with the accurate installation of concealed pipes, tiles, carpentry and other bathroom fittings. These activities are repetitive and often require additional manpower to help with using measurement, marking or lining tools. Also, the reference lines and markings are done separately by the different trades, and this could result in discrepancies.

In a concrete batching plant, it is important to measure the volume of cement that is entering the plant to plan and monitor operations. The conventional way of measuring cement volume is for the cement tanker to drive onto a weighbridge when entering and exiting the plant to calculate the difference in weight and derive the volume. The cement is enclosed in the tanker, so it is not possible to use visual means to determine the volume.