On the pointlessness of machine learning based time delayed prediction of TBM operational data

Erharter, Georg H.; Marcher, Thomas

doi:10.1016/j.autcon.2020.103443

Cited by 34 publications

(11 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This approach is therefore only able to capture the 'genuine patterns and relationships which exist in the historical data' [2]. That ML 'predicted sequences are only a slightly different version of the input sequence but shifted into the future' [1] should therefore come as no surprise.…”

Section: Based Forecastingmentioning

confidence: 99%

“…Three ML based forecasting techniques are evaluated, namely long shortterm memory (LSTM), support vector regression (SVR) and random forests (RF). The authors conclude that while 'high accuracy measures can be achieved, there is no predictive value within TDP forecasts' [1]. The purpose of this discussion is to make a case for a more optimistic outlook for the role of ML forecasting techniques in the tunnelling industry.…”

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

“…The recent paper by Erharter and Marcher [1] provides a timely and thought-provoking discussion of the use of machine learning (ML) for time delayed prediction (TDP) of operational data acquired from a tunnel boring machine (TBM). The stated goal of the Erharter and Marcher [1] study is 'to emphasize that ML based TDP for geotechnics is pointless'. To this end, the authors consider the specific problem of forecasting rock mass conditions ahead of the tunnel face using TBM operational data acquired from an Austrian construction site.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Discussion of “on the pointlessness of machine learning based time delayed prediction of TBM operational data” by Georg H. Erharter and Thomas Marcher

Sheil

2021

Automation in Construction

View full text Add to dashboard Cite

Section: Based Forecastingmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Discussion of “on the pointlessness of machine learning based time delayed prediction of TBM operational data” by Georg H. Erharter and Thomas Marcher

Sheil

2021

Automation in Construction

View full text Add to dashboard Cite

“…Existing estimation methods mainly apply probabilistic techniques to calculate an apartment frame construction period, which remains difficult because of the uncertainty 2 of 13 factors [5][6][7][8][9]. These methods increase estimation accuracy to a certain extent, but they are unlikely to be practically applied, owing to the difficulty in verifying ground truth estimations [10].…”

Section: Introductionmentioning

confidence: 99%

Quantitative Analysis of Waiting Length and Waiting Time for Frame Construction Work Activities Using a Queue Model; Focusing on Korean Apartment Construction

2021

View full text Add to dashboard Cite

The frame construction of an apartment complex that consists of multiple buildings encounters various uncertainties, owing to the complex relationships between units of work. Currently, the period of such a construction is calculated based on the number of floors of the highest building in the complex. This study quantitatively analyzes an apartment frame construction period using a queue model and evaluates the validity of the estimated period. In this regard, a methodology is proposed for analyzing the construction period by applying the concept of a customer and a server. A case study on the duration of an apartment frame construction period is conducted with the Korea Land and Housing Corporation, which has supplied the largest number of apartments in South Korea. It was found that the stable state of a queue system was observed when the rate of server utilization was applied to the basement and above-ground floors. However, a stable state was not reached on the ground floor. This study includes non-working days in its calculation and quantitatively analyzes uncertainty factors during construction. Therefore, the findings can be practically utilized to quantitatively plan the durations of work units in an apartment frame construction.

show abstract

Toward machine learning based decision support for pre‐grouting in hard rock

Rongved,

Hansen,

Erharter

2024

Civil Engineering Design

View full text Add to dashboard Cite

Pre‐grouting in hard rock tunneling is crucial for mitigating water ingress, significantly affecting project time and cost. Predicting pre‐grouting requirements is challenging and relies heavily on the expertise of on‐site personnel for decision‐making. This paper explores using supervised machine learning (ML) to create a data‐driven pre‐grouting decision process, aiming to predict “grouting time” and “total grout take.” Tree‐based regression models were developed using data from a Norwegian railway project, including typical tunneling data. These models showed limited predictive performance, with R2 scores of 0.40, though a significant relationship was observed. The limited performance highlights the need to identify parameters that significantly impact grouting outcomes rather than indicating the unsuitability of tree‐based models. Future research should consider a larger data set and additional parameters, such as more data on rock mass quality, hydrogeological conditions ahead of the face, and human, organizational, and contractual factors. Despite initial findings, supervised ML shows promise in enhancing data‐driven decision‐making in pre‐grouting by using appropriate input features and target variables.

show abstract

On the pointlessness of machine learning based time delayed prediction of TBM operational data

Cited by 34 publications

References 30 publications

Discussion of “on the pointlessness of machine learning based time delayed prediction of TBM operational data” by Georg H. Erharter and Thomas Marcher

Discussion of “on the pointlessness of machine learning based time delayed prediction of TBM operational data” by Georg H. Erharter and Thomas Marcher

Quantitative Analysis of Waiting Length and Waiting Time for Frame Construction Work Activities Using a Queue Model; Focusing on Korean Apartment Construction

Toward machine learning based decision support for pre‐grouting in hard rock

Contact Info

Product

Resources

About