Developing a Transmitter for High-Temperature, High-Speed Geothermal Data Transfer
The natural heat produced deep beneath the Earth’s surface is a clean source of energy that is potentially thousands of times greater than the world’s oil reserves. To access this energy source, engineers drill geothermal wells that are thousands of feet deep. To predict energy production rates, engineers then lower sensors (Figure 1) into the well to assess downhole conditions—including temperature, pressure, flow rates, acidity, and other characteristics.
Part of our team’s work in the Geothermal Research program at Sandia National Laboratories is in developing new technology for characterizing geothermal wells. We recently developed a data link for transmitting sensor data over a single-wire coaxial cable from deep within the well back to the surface. Working with a MathWorks Consulting Services engineer, we implemented the transmitter using MATLAB® and Communications Toolbox™, and then deployed it to a microcontroller rated for temperatures greater than 200 ˚C. The MathWorks engineer not only helped us optimize the transmitter design to maximize its data rate, but he also accelerated the deployment to our target microcontroller by generating code with Embedded Coder®.
Challenges in Downhole Instrumentation
The sensors used to measure the chemistry and temperature in geothermal wells produce relatively weak signals that cannot be reliably transmitted to the surface over thousands of feet of wire. One solution to this problem is to include a microcontroller, which can gather signals from multiple sensors and then transmit digital data using long-cable communication techniques. The choice of microcontroller, however, is constrained by the high temperatures that downhole instrumentation is exposed to. Few microcontrollers are designed to operate within the temperature range of geothermal wells.
In addition to anticipating the difficulties imposed by the depth and temperature of the well, we needed to design, implement, and test the data link rapidly to meet project schedule constraints.
Coding and Simulating the Transmitter
Before this project, the Geothermal department developed a QAM + OFDM communications protocol in MATLAB, implemented it using NI™ hardware to generate the signals, and developed a custom high-temperature line driver to drive signals on long cables [1]. This earlier effort yielded excellent data rate results over 5,000 feet of wireline but was not implemented on a high temperature processor. The next goal was to implement the code on a high-temperature microcontroller. We began modifying the code and adapted it to suit our requirements. However, due to pressing time constraints, we ultimately resolved to place MathWorks Consulting in charge of software implementation, while we concentrated on hardware development.
We met with a MathWorks engineer and gave him our requirements for the transmitter. There were four microcontrollers that we were considering that could withstand high temperatures, but after recommendation from the MathWorks engineer, we assessed a fifth microcontroller: a 32-bit high-temperature device from the Texas Instruments® C2000™ family. Ultimately, this device was chosen for the project due to simplified implementation. We proceeded to design and construct a custom printed circuit board (PCB) based on this microcontroller (Figure 2). Meanwhile, the MathWorks engineer updated and refined the transmitter code to achieve maximum efficiency for the protocol and hardware in these tests. To simplify the software for this implementation, binary phase-shift keying (BPSK) + OFDM was utilized instead of QAM + OFDM.
The process of optimization was carried out in a two-stage approach. First, the processing load needed to be trimmed to consume the lowest number of cycles while maximizing the data baud rate. In this application, it is acceptable to trade off latency for a higher baud rate. This tradeoff entails restructuring the architecture of the transmitter to achieve a higher level of optimization. In the second stage, the memory usage was fine-tuned. All these optimizations had to be done within the constraint of the chip’s capability. The optimized MATLAB code was disseminated to our team, and we evaluated its performance by conducting simulations and verifications locally.
Hardware Implementation and Testing
As the transmitter implementation neared completion, we asked the MathWorks engineer to deploy it to an evaluation board that included a low-temperature version of the Texas Instruments microcontroller. For this part of the project, he packed the MATLAB transmitter code in a Simulink® model and used Embedded Coder with Embedded Coder Support Package for TI C2000 to generate C code from the model. He then compiled the C code and executed it on the evaluation board to evaluate the real-time operation of the transmitter.
Once he had verified the implementation on the evaluation board, the engineer sent the model and generated code to us so that we could test on the custom PCB that we had built for the high-temperature microcontroller. We conducted tests by placing the PCB in an oven and analyzing the transmitted signals captured at the end of a 5,000-foot wire (Figure 3).
We conducted our initial tests at 170 ˚C and higher, capturing the transmitted signals using a Pico Technology oscilloscope at the recommendation of the MathWorks engineer, who used the same oscilloscope in his low-temperature testing on the evaluation board. We postprocessed and visualized the captured constellation data in MATLAB (Figure 4), finding that the data link operated successfully at a transfer rate of 30 kbps up to temperatures of 170 ˚C through 1,524 meters (5,000 ft) of high-strength high-temperature coaxial wireline. At temperatures higher than that, the amplifier utilized in the circuit degraded in performance and the signal was distorted. Removing the amplifier and wireline out of the equation, the microcontroller was successfully able to transmit data at 30 kbps to the oscilloscope at temperatures up to 210 ˚C.
Continuing Development
We initially presented our work on the data link at the 2022 Geothermal Rising Conference [2] and released a SAND report [3]reviewing the full project. Currently, we are looking ahead to the next phases of our research, in which we will update the design to handle higher temperatures with a new microcontroller rated for 300 ˚C, increase the constellation size to increase the data rates, reimplement QAM with dynamic constellation size modification, reimplement amplifier/line distortion correction, and implement error correction. Due to the positive experience we had working with MathWorks Consulting Services, we plan to work with the same engineer on this new project.
Acknowledgments
This effort was funded by the U.S. Department of Energy, Office of Energy Efficiency and Renewable Energy, and Geothermal Technologies Program.
We gratefully acknowledge the exceptional contributions of MathWorks Consulting Services Engineer Francis Tiong in the rapid development, optimization, and deployment of the data link software.
Funding Statement
This paper describes objective technical results and analysis. Any subjective views or opinions that might be expressed in the paper do not necessarily represent the views of the U.S. Department of Energy or the United States Government.
Sandia National Laboratories is a multimission laboratory managed and operated by National Technology & Engineering Solutions of Sandia, LLC, a wholly owned subsidiary of Honeywell International Inc., for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-NA0003525. (SAND2023-03536N)
References
[1] Cashion, Avery, and Grzegorz Cieslewski, “High Temperature Quadrature Amplitude Modulation over Orthogonal Frequency Division Multiplexing.” IMAPSource Proceedings 2017 (HiTEN): 20–30. https://doi.org/10.4071/2380-4491.2017.HiTEN.20
[2] Wright, Andrew A., Avery Cashion, and Francis Tiong. “High Temperature High Speed Downhole Data Transfer (Data Link).” GRC 2022 Conference. Reno, Nevada, 2022.
[3] Wright, Andrew A., Avery Cashion, and Francis Tiong. “High Temperature Component and Data Link Evaluation.” SAND Report (February 2023). SAND2023-00068.