OEDI SI/Scenarios/Extended Kalman Filter DSSE

From Open Energy Information

Extended Kalman Filter DSSE​ Summary

  • Date Created: 2024/01/03
  • Organization: PNNL
  • Objective: Distribution system state estimation (DSSE) refers to the process of estimating the current conditions of the system’s state, such as voltage/current magnitudes and angles, power flows, etc. in a distribution system using available field measurements from sensors installed at substations, feeders, and other strategic points in the distribution network. Since the distribution system is inherently unbalanced and three-phase in nature, the traditional state estimation for transmission systems cannot be readily implemented for distribution systems. This DSSE process is essential for monitoring and controlling the unbalanced distribution network. By accurately estimating the system state, operators can detect abnormalities, identify potential issues, and optimize the operation of the distribution grid. For this developed DSSE, the Extended Kalman Filter (EKF) -based method is adopted to estimate the system state considering historical estimates and any new measurements. The EKF is an extension of the standard Kalman Filter, designed for estimating states for the nonlinear systems by linearizing them around the current estimate. Generally, the EKF operates by predicting the system's state forward in time using a dynamic model and incorporating measurements to refine the estimate. Unlike the standard Kalman Filter, which assumes linear dynamics and measurements, the EKF accommodates nonlinearity by approximating it using first-order Taylor series expansion. The key component of the EKF is the Jacobian matrix, which represents the linearization of the system dynamics and measurement functions at the current state. Specifically, in the EKF algorithm, a state transition model remembers the previous state estimate and introduces additional uncertainty according to the time between estimates. No specific set of new measurements is required at any time step. The EKF algorithm is computationally efficient, using a single iteration to take a near-optimal step towards the maximum-likelihood state estimate.
  • Use Case: Distribution System State Estimation
  • Methodology
    • Inputs
      • Outputs
        • Configuration
          • Webinars


          Docker Container

          Download the docker container at:


          Run Locally

          http://localhost:8080/edit_scenario?Extended Kalman Filter DSSE
              Run Locally


          Input Data

          Download input data at:


            Output Data

            Download output data at:

              Component Raw Data

              See Component Raw Data:
              IEEE 123-bus Distribution Network Data

              References


                Back to Distribution System State Estimation



                Property "Objective" (as page type) with input value "Distribution system state estimation (DSSE) refers to the process of estimating the current conditions of the system’s state, such as voltage/current magnitudes and angles, power flows, etc. in a distribution system using available field measurements from sensors installed at substations, feeders, and other strategic points in the distribution network. Since the distribution system is inherently unbalanced and three-phase in nature, the traditional state estimation for transmission systems cannot be readily implemented for distribution systems. This DSSE process is essential for monitoring and controlling the unbalanced distribution network. By accurately estimating the system state, operators can detect abnormalities, identify potential issues, and optimize the operation of the distribution grid. For this developed DSSE, the Extended Kalman Filter (EKF) -based method is adopted to estimate the system state considering historical estimates and any new measurements. The EKF is an extension of the standard Kalman Filter, designed for estimating states for the nonlinear systems by linearizing them around the current estimate. Generally, the EKF operates by predicting the system's state forward in time using a dynamic model and incorporating measurements to refine the estimate. Unlike the standard Kalman Filter, which assumes linear dynamics and measurements, the EKF accommodates nonlinearity by approximating it using first-order Taylor series expansion. The key component of the EKF is the Jacobian matrix, which represents the linearization of the system dynamics and measurement functions at the current state. Specifically, in the EKF algorithm, a state transition model remembers the previous state estimate and introduces additional uncertainty according to the time between estimates. No specific set of new measurements is required at any time step. The EKF algorithm is computationally efficient, using a single iteration to take a near-optimal step towards the maximum-likelihood state estimate." contains invalid characters or is incomplete and therefore can cause unexpected results during a query or annotation process.


              1. This scenario employs an extended Kalman filter (EKF) based method for DSSE. The EKF method has two-steps i.e., prediction step and an update step. Note that the voltage magnitudes and voltage angles are the states to be estimated. Therefore, the number of states is twice as many compared to the total number of nodes in the system.
              2. The DSSE algorithm is integrated in OEDI SI as an independent federate, where HELICS maintains a message queue. This allows each federate to move at their own pace. The use case represents a time series analysis at 15-minute intervals for 24 hours. The sensor federate generates the measurements using the power flow results generated by the power flow federate. At each time step, the EKF algorithm (DSSE) federate generates the voltage magnitude and angle estimates by using the measurement set and topology data. Results are logged by the recorder federate as well as written in the .csv files. They are also plotted using the post-processing scripts.
              3. Node names, nominal node voltages, and angles
              4. System Y-bus matrix
              5. Location of source bus
              6. Nominal active and reactive power loads at all nodes (used for pseudo-measurements)
              7. Measurements of voltage magnitudes
              8. Measurements of real and reactive powers
              9. Location of all measurements. Note that the measurements are randomly generated at 20 % of the total nodes (as configured in the sensor federate). For the rest of the nodes, DSSE federate generates the pseudo-measurements.
              10. Estimated voltage and the estimated angle at all nodes
              11. The user is not required to manipulate the internal contents of this image. To run the image, the user needs to follow the instructions in the readme file.
              12. https://hub.docker.com/r/openenergydatainitiative/pnnl-dsse-ekf
              13. https://data.openei.org/s3_viewer?bucket=oedi-data-lake&prefix=SMART-DS%2Fv1.0%2F
              14. https://data.openei.org/s3_viewer?bucket=gadal&prefix=gadal_ieee123%2F
              15. Output are saved to the local folder /outputs or the designated mounted volume for the single container setup.
              16. https://github.com/GRIDAPPSD/gridappsd-state-estimator/tree/OEDISI.1.1
              17. F. B. dos Reis, A. P. Reiman, and G. D. Black, “Distributed multi-area state estimation for distribution systems,” in 2022 IEEE Power Energy Society General Meeting (PESGM), 2022, pp. 1–5. https://ieeexplore.ieee.org/document/9917205.
              18. https://data.openei.org/submissions/5915
                19. Retrieved from "https://openei.org/w/index.php?title=OEDI_SI/Scenarios/Extended_Kalman_Filter_DSSE&oldid=1084927"