Power systems equipment technical condition analysis and diagnostics
Power system equipment condition monitoring is a crucial task for ensuring electricity supply reliability and efficiency. The major goals are usually achieved by equipment maintenance and replacement, which is being performed based on typical service life parameters. The way from the rule-based to the model-based monitoring and diagnostics brings significant economic benefits.
Our solution currently consists of:
Condition-based data recording (faults, abnormal conditions, fixed time records)
Risk assessment and consequences evaluation
Flexible conditions alarming and intelligent prediction
Adaptive equipment and processes digital models
Machine learning based pattern recognition
Statistics accumulation and analysis
Instantaneous electrical parameters
Estimation of magnitude, frequency and phase values of steady-state electric parameters (current and voltage) is an urgent problem nowadays for dynamic processes study in non-linear systems (in power systems in general as well as in generators particularly).
Modern phasor measurement units provide measurements of 50/60 Hz sampling rate (one measurement per cycle) according to the IEEE С37.118-2005 standard. To date there is a number of methods employed for state parameters calculation. These techniques are based on the assumption of signal parameters permanence.
It is in conflict with non-linear characteristics of electrical systems equipment. Existing approaches to state parameters estimation with sampling rate equal to fundamental frequency don’t allow to precisely calculate state parameters values at the arbitrary time points during transients.
Measurements data validation
The measurement data validation system is proposed for the parameters measurements based on the optimization techniques and progressive hardware solutions. The following hierarchical structure is considered:
1st level is responsible for validating the initial measurements raw data recorded by the PMUs and the DFRs within one power system entity (power plant, substation, microgrid).
2nd level provides data validation involving the neighboring objects.
3rd level corresponds to the control centers.
The implementation of all levels is possible as a single software solutions and as separate systems as well. This methodology implies the power system model parameters are predefined or imported from the existing models, but are rectified during operation on the basis of collected data.
Power systems equipment digital twins
Load and generation simulation are highly relevant: the former lacks solid theoretical foundation, and numerous approaches are applied depending upon the considered conditions; the latter involves complex models associated with heavy computational burden and parameters definition challenge.
For the power system simulation purposes any load might be represented by its frequency and voltage response. The method is proposed for defining the frequency static response based on synchrophasor measurements during transients accompanied by frequency deviations. The method was successfully validated involving four events in a power system region resulting in frequency deviations of up to 0.06 Hz magnitude and recorded by means of WAMS.
Moreover, synchrophasors allow calculating generation electrical parameters during electromechanical disturbances, hence one can use simplified generating unit models, parameters of which can be calculated using field measurements. In particular, generating units damping properties might be estimated. A method is proposed for defining the synchronous machine damping power and the inertia constant of a generating unit based on synchrophasor measurements analysis and the generator rotor swing equation. These parameters allow determining the generating unit capability to damp power system oscillations. The method was tested in MATLAB Simulink and with physical simulator.
Advanced IT and data analytics
Modern electrical grids together with renewables demand the two way communication to continuously sense, analyze, pre-empt, and act upon potential uncertainties. Electrical grid of this scale would require unimaginable amounts of data and processing power, which would be challenging with our existing network setup. Managing the collected data would require huge data centers with advanced data analytical software. The only solution to address this issue and reduce maintenance cost is distributed, cloud computing. We need to transfigure the canonical hierarchical structure to benefit the system in terms of both functionality and reliability.
We suggest gaining benefit from the edge computing paradigm, which implies putting computing power closer to the source of data meeting the need for processing and analytics for sensor data. This is the proper response to the data handling challenge.
Another upcoming trend in line with edge computing is controllers virtualization, i.e. replacing several physical devices by several virtual instances operating on a same hardware platform, be it cloud-based or on-premises installation. It significantly increases the operation reliability and decreases deployment and maintenance expenses.
Relay protection and automatic emergency control management
The development is underway of a complete solution for the relay protection and automatic emergency control systems management, which aims at addressing the whole range of challenges - from settings evaluation and adjustment and up to risk-based control. The solution is designed for various scale energy industry entities to minimize outages impact and conduct preventive diagnostics.