• Wrote a custom environment to implement Deep Reinforcement Learning algorithms (DDPG, SAC, PPO, MADDPG, MASAC) together with attention-based forecasting recurrent neural network models for operation of a set of utility scale batteries resulting in a novel approach for battery action prediction. Utilized Keras and PyTorch for algorithms, gym environment base class for environment construction, as well as MongoDB and TensorBoard for storage.

• Created synthetic ampacity and power price data with Wasserstein Generative Adversarial Networks for applications in battery capacity sizing resulting in more accurate battery capacity calculations.

• Trained a team of energy management engineers and ensured efficient real time operations of company's generating assets by executing protocols as per AESO rules which resulted in minimal compliance violations and improved relations with stakeholders.

• Worked cross-functionally and build relations with interval and external commercial management, engineering, regulatory, and real-time operations teams to execute projects in the 0.1-1M$ range resulting in revenue clawbacks and installations of new powerplant equipment on budget and on time.

• Implemented business processes and team culture improvements by creating macro applications using VBA and python programming languages, and implementing data storage solutions using MySQL, resulting in improvements in efficiency and reduced number of compliance reports.

• My goal is to apply algorithms to data and utilize my interpersonal skills to solve complex real-world problems.

---

Energy Management Engineer / Capital Power

2013 - 2016, 2017 - 2019

• Monitored generation of power assets for compliance with AESO Rules parts 100-500 and AESO’s Operating Policies and Procedures using dashboards compiling PI and SCADA data as part of Energy Management Operations Center’s (EMOC’s) operations
• Managed, trained, and mentored new hires on every aspect of energy management ensuring efficient control room operations; utilized JIRA to perform User Acceptance Testing and suggested improvements to proprietary business applications resulting in improved efficiency and ensuring reliable performance of software
• Leveraged C# to analyze proprietary software performance in accordance with business logic and implemented improvements to source codes resulting in correct offer aggregation; utilized SQL to compile data from various sources to create intuitive reports for executive management; created VBA-based applications which resulted in efficient execution of daily working processes and timely delivery of reports
• Used PIProcessBook and PowerBI to create the display layout and determined necessary variables to assist operations in determining the status of power plant operations; utilized Python, R, and proprietary software to analyze wind asset data resulting in increased wind asset utilization
• Worked with commercial management, AESO, and compliance stakeholders to lead initiative to capture missing revenue from Ancillary Service provision resulting in the claw-back of payment
• Created coal allocation model tracking the quality of coal in coal slots to assist operations in identifying the relationships between flue gas emissions and coal cuts; determined the impact of AESO’s dynamic setpoint and the resulting alarms on a wind site resulting in the identification of revenue losses
• Re-examined settlement results and re-created complex settlement procedures for the ESA contract resulting in the implementation of a more efficient settlement engine
• Investigated multiple start-up scenarios resulting in reduced ratchet charges for the sequence of generating plants coming online; led the South Power Pool market review to identify EMOC’s feasibility of service provision in the United States energy markets
• Coordinated and scheduled real-time outages and testing of EMOC’s assets resulting in compliance with regulatory standards and procedures of AESO, BC Hydro, IESO, Hydro One, South Power Pool, and MISO
• Performed restatements on the operating characteristics of power-generating assets resulting in compliance with operating standards of local jurisdictions
• Analyzed and submitted energy offers ensuring efficient pricing structuring of assets for Capital Power’s stakeholders; created, edited, reviewed, and organized working-level operating procedures resulting in more efficient real-time performance
• Reviewed technical specifications of fossil-fuel assets resulting in creation of efficient working-level procedures; created, reviewed compliance violation reports, and drafted responses to Market Surveillance Administrator ensuring EMOC maintains its reputation of a reliable control center

Mechanical Engineer / Capital Power

2016 - 2017

• Managed and executed various projects ranging from $0.1M - $1M including oxidization skid upgrade, drag chain conveyor replacement, and coal dust collector installation
• Sourced vendors, created budgets, evaluated RFQs, prepared documentation, and created mechanical designs for power plant equipment using Autodesk software suite
• Coordinated scaffolding, mechanical, insulation, and Non-destructive Examination (NDE) contractors to inspect and repair welds ensuring power plant outages adhere to schedule and budget
• Led status meetings and prepared daily progress reports on the completion of weld inspections to the turnaround outage group; Checked mechanical drawings and P&IDs for compliance with design intent
• Performed calculations on weld thicknesses based on NDE reports ensuring pipe thickness meets specifications
• Attended status meetings and prepared daily progress reports on the completion of weld inspections
• Tracked the cost of NDE contractor labor ensuring adherence to budget; communicated with engineering and reliability teams to complete any missing work ensuring set budgets were followed

Mechanical Engineer / Academy Petroleum Industries

2011

Used SolidWorks, Autodesk Inventor, and AutoCAD to design oil-well equipment in compliance with API 6A, 6B, and 16C standards; checked drawings to comply with code and suggested improvements to document-handling process that resulted in faster access to archived materials.

• Battery action prediction using MILP and RNN-forecasting.

• Environment for 6-bus test grid.

• DDPG applied to battery control in 6 bus grid.

• MASAC applied to battery control in 6 bus grid.

• SAC applied to battery control in 6 bus grid.

• MADDPG applied to battery control in 6 bus grid.

• WGAN for synthetic time ampacity and electricity pool prices.

• CWGAN-GP for mnist digits generation (link).

• CWGAN-GP for synthetic time series (link).

• CNN-BiLSTM-with-attention for load forecasting on 24-bus RTS.

• Diet calculator that minimizes food cost (link).

• Battery sizing using non-linear programming taking into account transmission line outages, battery degradation, and dynamic line rating applied to 24-bus RTS.

• Method for creating multi-bus battery sizing simulation.

• Multiprocessing price scraper for a selected cryptocurrency to find price discrepancies, notify user, and perform abritrage trades utilizing Twillio api (link).

• Cryptocurrency price scraper using requests and json libraries (link).

• Live dashboard for Alberta Electric System Operator (AESO) Pool Price forecast and comparison against AESO forecast.

• Links to projects coming soon! For now please visit https://www.github.com/avkhimen/.

---