2018 Industrial Problem Solving Workshop (IPSW)

Mentors

• Juris Meija (NRC Metrology)

Preparation

• Statistics
• Experimental Design
• Optimal Design
• Data Analysis

Overview

Certified reference materials (CRMs) are key to global comparability of measurement results as most laboratory measurements calibrate or verify their measurement results against the international CRMs. CRMs are typically created in a batch of several thousand units and it is inevitable that the property values (such as the mass fraction of arsenic) might vary between the units. This variability is captured as the uncertainty due to homogeneity of the material and is part of the overall combined uncertainty of the certified values. The question of how many samples need to be tested in order to get a reliable estimate of the homogeneity uncertainty is raised here. This question has financial as well as technical implications. If too few samples are analyzed, one might underestimate the homogeneity of the CRM which in turn might lead to diminished trust in future CRMs or a possible recall. Analyzing too many samples, on the other hand, is undesirable as it requires unnecessary expenses, labour, and depletion of the CRM stock.

Download the full project description (pdf)

Mentors

• Nathan Vadeboncoeur (Smart Shores)

Preparation

• Approximations of functions
• Signal processing
• Image processing
• Statistics
• Data science



Smart Shores is a science-based tech company focused on making world-class environmental data accessible to everyone, from small not-for-profits to big governments. We provide afford- able high-resolution spatial data, ecological analysis and visual media to support responsible and effective stewardship of our natural world.

 

We help tell the story behind the data by providing both technical (GIS) data and engag- ing content including interactive 3D environments, detailed maps, short videos, and outreach materials.

 

A key focus for our many of our clients is environmental management, which is crucial for our quality of life. For example, a well-managed shoreline can support salmon and shorebird populations, provide recreation opportunities and buffer against flooding. However, shore- line development can also have unintended consequences that are harmful to ecosystems, the economy, recreation and culture. We help connect human and ecological needs to support progressive, twenty-first century land management.




Download the full project description (pdf)

 

 

Mentors

• Thomas Holloway (Responsive AI)

Preparation

• Data Science
• Statistics
• Statistical Modelling
• Some Finance

Overview

Responsive AI is a venture-backed startup in direct-to-client wealth management, hybrid wealth technology (B2B), and data-driven client intelligence research. This summer we are celebrating three years of operation and the close of our US $1.1mm seed financing round.

Responsive’s asset management business is differentiated by streamlined user experience, industry-leading cost, and the implementation of data science research for optimized portfolio performance. Eschewing security selection as a source of value, Responsive clients own portfolios of Exchange Traded Funds (ETFs). Each ETF provides exposure to an asset class such as US stocks, foreign stocks, Government bonds, and High Yield Bonds. The Responsive investment process is to reconsider the weights monthly using signals from markets and the economy. Following Samuelson, “The market is micro efficient, but macro inefficient”.

Download the full project description (pdf)

Mentors

• Matt McDonald (Fotech Solutions Ltd.)

Preparation

• Calculus and Mathematical Modelling
• Numerical Methods
• Signal Processing
• Some physics background would be useful

Overview

Distributed acoustic sensing (DAS) is a relatively new technology used for measuring strain and vibration along a long cable installed on large linear infrastructures such as railways, roadways, pipelines, underground boreholes, and so forth. The sensor consists of a single-mode fibre optic cable installed along the length of structure under surveillance, and connected to a laser and photodetector. A pulse of light is sent down the fibre, and the reflections scattered back to the detector are analyzed to measure strain on the fibre. Depending on the length of the fibre, the laser may be fired many thousands of times per second, and the Rayleigh backscatter sampled to produce a measurement of the strain at numerous points on the fibre. Knowledge of the refractive index of the fibre allows one to use two-way travel-time to treat the fibre as a series of discrete sensors up to 40 km long, where each position acts somewhat like a two-beam interferometer.

In terms of typical data volumes collected, a 5 km fibre with a spatial sampling interval of 0.67 meters would translate to roughly 7462 fibre-positions being sampled 20,000 times per second. The signals can be treated as the output of a long string of individual sensors distributed along the cable, with a high frequency sampling rate.

A significant challenge in improving the sensitivity and accuracy of the DAS device is calcu- lating the optical phase of the light signal travelling in the fibre from the measured intensities on the photodetector. Noise and phase wrapping are the main issues that make this difficult.

Download the full project description (pdf)

Mentors

• Michael Morgan (GLJ Petroleum Consultants)

Preparation

• Differential equations
• Stochastic modelling
• Statistical models
• Time series analysis
• Econometrics
• Numerical methods
• Some knowledge of financial mathematics and asset pricing would be great

Many aspects of the oil and gas industry are subject to economic uncertainty: success in finding oil and gas reserves is highly unpredictable, productivity of an individual well can change in seemingly random ways, markets for products can move unexpectedly, future prices and interest rates are unknown, among other challenges. Useful economic models need to account for these uncertainties. Stochastic models provide an approach to these problems by treating certain unknown parameters as random variables and performing an in-depth analysis that takes into account this uncertainty.

Stochastic differential equations (SDEs) are one useful method whose application are of interest to oil and gas economic problems. As most economic analysis is currently conducted using deterministic discounted cash flow models, there is much scope of innovation and experi- mentation. In particular, there is much scope for improving how uncertainty is mathematically modelled. In these project, we are interested four different general areas:

• Econometrics
• Real Options Analysis
• Geographic Correlation and Temporal Uncertainty
• SDE Solutions to PDEs

Download the full project description (pdf)