Jump to content

Pandapower Architecture

From Game in the Brain Wiki

Pandapower Architecture & Codebase Guide

Template:Quote

1. What Is Pandapower?

Pandapower is a Python library for power system modeling and analysis. It lets you:

  • Build electrical network models (buses, lines, transformers, generators, loads, etc.)
  • Run simulations (power flow, optimal power flow, short-circuit, state estimation, time series)
  • Visualize results (static plots, interactive Plotly maps)
  • Import/export networks (JSON, Excel, SQL, MATPOWER, CIM, PowerFactory)

Everything is built on pandas DataFrames — each network element type is a table, indexed by element ID. This makes it straightforward to serialize to/from JSON for a web API.

2. Project Root Layout

pandapower/
├── pandapower/          # Main Python package (the library itself)
├── tutorials/           # 60 Jupyter notebooks — examples & learning material
├── doc/                 # Sphinx RST documentation
├── .github/             # CI/CD workflows
├── pyproject.toml       # Dependencies, build config, Python 3.10–3.14
├── setup.py             # Minimal setuptools entry
├── README.rst
├── CHANGELOG.rst
├── AUTHORS
├── CITATION.bib
├── architecture.md      # ← This file
└── GUI_PRD.md           # Functional requirements for the web GUI

3. Core Package Structure (pandapower/pandapower/)

3.1 Network Creation & Data Model

Path Purpose
create/ Element creation functions — one file per element type
create/__init__.py Re-exports all create_*() functions
create/bus_create.py create_bus(), create_buses()
create/line_create.py create_line(), create_line_from_parameters()
create/gen_create.py create_gen(), create_gens()
create/load_create.py create_load(), create_loads()
create/sgen_create.py create_sgen() — static generators (solar/wind)
create/ext_grid_create.py create_ext_grid() — slack/reference bus
create/impedance_create.py create_impedance() — series impedance
create/shunt_create.py create_shunt() — capacitor banks
create/switch_create.py create_switch() — breakers/isolators
create/storage_create.py create_storage() — batteries
create/motor_create.py create_motor() — async motors
create/cost_create.py create_poly_cost(), create_pwl_cost() — OPF costs
create/measurement_create.py create_measurement() — for state estimation
create/network_create.py create_empty_network() — the starting point
create/group_create.py create_group() — element grouping
create/source_create.py DC elements, FACTS (SVC, TCSC, SSC, VSC)

Key concept: A pandapowerNet is a dict-like object where each key (e.g., net.bus, net.line, net.load) is a pandas DataFrame. Results go into net.res_bus, net.res_line, etc.

3.2 Analysis Engines

Path Function What It Does
run.py runpp() AC Power Flow — the most-used function
run.py rundcpp() DC Power Flow (linearized approximation)
run.py runopp() AC Optimal Power Flow
run.py rundcopp() DC Optimal Power Flow
run.py runpp_3ph() Three-phase unbalanced power flow
run.py runpp_pgm() Power flow via C++ PowerGridModel backend
powerflow.py Internal Power flow orchestration
optimal_powerflow.py Internal OPF orchestration
pf/ Internal Newton-Raphson, backward-forward sweep, Gauss-Seidel, fast-decoupled solvers
shortcircuit/ calc_sc() Short-circuit analysis (IEC 60909)
estimation/ estimate() State estimation (WLS, IRWLS, LP)
timeseries/ run_time_series() Multi-timestep simulation
contingency/ run_contingency() N-1 contingency analysis
opf/ Internal OPF via PYPOWER / PowerModels.jl

3.3 Results

Path Purpose
results.py Main result extraction orchestrator
results_bus.py Bus results: voltage magnitude/angle, P/Q injection
results_branch.py Line & trafo results: loading%, current, losses
results_gen.py Generator results: P/Q output

Result tables populated after analysis:

Table Key Columns
net.res_bus vm_pu, va_degree, p_mw, q_mvar
net.res_line loading_percent, i_from_ka, i_to_ka, pl_mw, ql_mvar
net.res_trafo loading_percent, pl_mw, ql_mvar
net.res_load p_mw, q_mvar
net.res_gen p_mw, q_mvar, vm_pu
net.res_sgen p_mw, q_mvar
net.res_bus_sc ikss_ka, ip_ka, ith_ka (short-circuit)
net.res_bus_est vm_pu, va_degree (state estimation)

3.4 I/O & Converters

Path Purpose
file_io.py to_json(), from_json(), to_excel(), from_excel()
sql_io.py to_sql(), from_sql() — PostgreSQL, SQLite
io_utils.py Serialization helpers
converter/matpower/ MATPOWER format import/export
converter/pypower/ PYPOWER format
converter/cim/ IEC Common Information Model
converter/powerfactory/ DIgSILENT PowerFactory
converter/ucte/ UCTE exchange format
converter/jao/ JAO format
converter/pandamodels/ Julia PowerModels.jl bridge

3.5 Visualization & Plotting

Path Purpose
plotting/simple_plot.py simple_plot() — Matplotlib static plots
plotting/collections.py Styled element collections (buses, lines, trafos)
plotting/powerflow_results.py Color-coded result overlays
plotting/geo.py Geographic coordinate handling
plotting/to_html.py Export to standalone HTML
plotting/plotly/ Interactive Plotly plots (most relevant for web GUI)
plotting/plotly/simple_plotly.py simple_plotly() — interactive network diagram
plotting/plotly/mapbox_plot.py mapbox_plot() — geographic map overlay
plotting/plotly/pf_res_plotly.py pf_res_plotly() — interactive results

3.6 Topology & Diagnostics

Path Purpose
topology/create_graph.py Convert network to NetworkX graph
topology/graph_searches.py Connected components, shortest paths, unsupplied buses
diagnostic/ Network validation: isolated buses, duplicate lines, impossible configs

3.7 Advanced Features

Path Purpose
control/ Controller framework — tap changers, DER, shunt, PQ control
protection/ Protection device modeling — fuses, relays
grid_equivalents/ Network reduction (REI, Ward equivalents)
groups.py Batch operations on element groups
std_types.py Standard component library (line types, trafo types)
toolbox/ Grid modification, element selection, data manipulation

3.8 Internal Infrastructure

Path Purpose
__init__.py Public API — imports everything users need
auxiliary.py Helpers, ADict class, logging
build_bus.py Internal bus matrix construction
build_branch.py Internal branch matrix construction
build_gen.py Internal generator matrix construction
pd2ppc.py Convert pandapower → PYPOWER internal format
convert_format.py Version migration between pandapower versions
network_schema/ Pandera DataFrame validation schemas
pypower/ Embedded PYPOWER solver algorithms

4. Built-In Test Networks (pandapower/networks/)

These are ready-made networks useful for testing the GUI:

Function Description Size
simple_four_bus_system() Tiny test network 4 buses
panda_four_load_branch() Small LV feeder ~10 buses
case_ieee30() IEEE 30-bus test case 30 buses
case118() IEEE 118-bus test case 118 buses
case300() IEEE 300-bus test case 300 buses
mv_oberrhein() Real German MV grid ~300 buses
create_cigre_network_mv() CIGRE MV benchmark ~15 buses
create_kerber_*() Synthetic LV grids Various

5. Tutorials (tutorials/)

60 Jupyter notebooks organized by topic:

Category Key Notebooks
Getting Started minimal_example.ipynb, create_simple.ipynb, create_advanced.ipynb
Power Flow powerflow.ipynb, convergence_powerflow.ipynb, three_phase_loadflow_tutorial_simple.ipynb
OPF opf_basic.ipynb, opf_curtail.ipynb, opf_dcline.ipynb
Short Circuit shortcircuit.ipynb
State Estimation state_estimation.ipynb
Time Series time_series.ipynb, time_series_advanced_output.ipynb
Contingency contingency_analysis.ipynb
Visualization plotting_basic.ipynb, plotly_built-in.ipynb, plotly_maps.ipynb
Controls control_loop.ipynb, building_a_controller.ipynb, DER_control_tutorial.ipynb
Protection protection/fuse.ipynb, protection/oc_relay.ipynb
Converters cim2pp.ipynb, converter_powerfactory.ipynb

6. Dependencies

Core (always required):

  • pandas ~= 2.3 — DataFrames (the foundation)
  • numpy >= 1.26 — Numerical arrays
  • scipy < 1.17 — Sparse matrices, solvers
  • networkx ~= 3.4 — Graph/topology algorithms
  • tqdm ~= 4.67 — Progress bars
  • geojson ~= 3.2 — GeoJSON support
  • pandera ~= 0.26 — DataFrame validation

Visualization (relevant for GUI):

  • plotly ~= 6.3 — Interactive web-based plots
  • matplotlib ~= 3.10 — Static plots
  • geopandas ~= 1.1 — Geographic data

Performance:

  • numba ~= 0.61 — JIT-compiled solvers
  • lightsim2grid ~= 0.12 — C++ power flow backend

I/O:

  • xlsxwriter, openpyxl — Excel
  • psycopg ~= 3.2 — PostgreSQL
  • lxml ~= 6.0 — XML parsing

7. How Pandapower Works (The Computation Flow)

┌─────────────────────────────────────────────────────┐
│  1. CREATE NETWORK                                  │
│     net = pp.create_empty_network()                 │
│     pp.create_bus(net, vn_kv=110)                   │
│     pp.create_line(net, from_bus=0, to_bus=1, ...)  │
│     pp.create_load(net, bus=1, p_mw=10)             │
│     ...                                             │
│     → net.bus, net.line, net.load are DataFrames    │
└───────────────────────┬─────────────────────────────┘
                        │
                        ▼
┌─────────────────────────────────────────────────────┐
│  2. RUN ANALYSIS                                    │
│     pp.runpp(net)           # Power flow            │
│     pp.runopp(net)          # Optimal power flow    │
│     pp.shortcircuit.calc_sc(net)  # Short circuit   │
│     pp.estimation.estimate(net)   # State est.      │
│                                                     │
│     → Internally: DataFrames → sparse matrices →    │
│       solver → results back to DataFrames           │
└───────────────────────┬─────────────────────────────┘
                        │
                        ▼
┌─────────────────────────────────────────────────────┐
│  3. READ RESULTS                                    │
│     net.res_bus      → voltage, power at each bus   │
│     net.res_line     → loading, current, losses     │
│     net.res_trafo    → loading, losses              │
│     net.converged    → True/False                   │
│                                                     │
│     → All results are pandas DataFrames             │
└───────────────────────┬─────────────────────────────┘
                        │
                        ▼
┌─────────────────────────────────────────────────────┐
│  4. VISUALIZE / EXPORT                              │
│     pp.plotting.simple_plotly(net)   # Web plot     │
│     pp.to_json(net)                  # Serialize    │
│     pp.to_excel(net, "network.xlsx") # Export       │
└─────────────────────────────────────────────────────┘

8. Key Architecture Facts for GUI Developers

8.1 Everything Is a DataFrame

  • Input data: net.bus, net.line, net.trafo, net.load, net.gen, net.sgen, net.ext_grid, net.switch, net.shunt, net.storage, etc.
  • Result data: net.res_bus, net.res_line, net.res_trafo, etc.
  • Both are pandas DataFrames — trivially convertible to JSON via .to_dict(orient="records") or .to_json().

8.2 JSON Serialization Is Built In

json_str = pp.to_json(net)       # Network → JSON string
net = pp.from_json(json_str)     # JSON string → Network

This is the natural transport format for a REST API.

8.3 Analysis Is Stateless (Almost)

  • Call runpp(net) → results appear in net.res_* tables
  • Call it again → previous results are overwritten
  • The net object carries all state — pass it around as the single source of truth

8.4 Plotly Is Already Web-Ready

  • pandapower.plotting.plotly generates Plotly figures
  • Plotly figures serialize to JSON natively
  • They can be rendered directly in a browser via plotly.js

8.5 Standard Types Are a Built-In Library

  • net.std_types["line"] — DataFrame of line types with impedance/capacity data
  • net.std_types["trafo"] — DataFrame of transformer types
  • When creating elements, pass std_type="NAYY 4x50 SE" to auto-populate parameters
  • GUI should present these as dropdowns

8.6 Controllers Run Inside Power Flow

  • runpp(net, run_control=True) activates the controller loop
  • Controllers (tap changers, DER control) iterate until convergence
  • For the GUI: expose controller configuration as a settings panel

For a Python-backend web GUI wrapping pandapower: 

┌──────────────────────────────────────┐
│  Frontend (Browser)                  │
│  - React / Vue / Svelte              │
│  - Plotly.js for network diagrams    │
│  - AG Grid or similar for tables     │
│  - PDF generation (jsPDF / pdfmake)  │
└──────────────┬───────────────────────┘
               │ REST / WebSocket
               ▼
┌──────────────────────────────────────┐
│  Backend (Python)                    │
│  - FastAPI or Flask                  │
│  - pandapower as the engine          │
│  - Endpoints for CRUD + analysis     │
│  - Session/project management        │
└──────────────────────────────────────┘

API Design Sketch:

Endpoint Method Purpose
/api/network POST Create new empty network
/api/network GET Get current network as JSON
/api/network/import POST Import from JSON/Excel/MATPOWER
/api/network/export GET Export to JSON/Excel
/api/element/{type} POST Create element (bus, line, load...)
/api/element/{type}/{id} PUT Update element parameters
/api/element/{type}/{id} DELETE Remove element
/api/analysis/powerflow POST Run power flow
/api/analysis/opf POST Run optimal power flow
/api/analysis/shortcircuit POST Run short-circuit analysis
/api/analysis/estimation POST Run state estimation
/api/analysis/timeseries POST Run time series
/api/analysis/contingency POST Run contingency analysis
/api/results/{type} GET Get result tables
/api/plot/network GET Get Plotly figure JSON
/api/report/{type} GET Generate PDF report
/api/stdtypes/{element} GET List standard types
/api/diagnostic POST Run network diagnostics
/api/topology GET Get topology info

10. Quick Reference: Most Important Files

If you're building a GUI and need to understand pandapower, read these files in order:

  1. pandapower/__init__.py — See every public function
  2. pandapower/run.py — Power flow & OPF entry points
  3. pandapower/create/__init__.py — All element creation functions
  4. pandapower/file_io.py — JSON/Excel serialization
  5. pandapower/results.py — How results are extracted
  6. pandapower/plotting/plotly/simple_plotly.py — Interactive visualization
  7. pandapower/shortcircuit/calc_sc.py — Short-circuit entry point
  8. pandapower/estimation/state_estimation.py — State estimation entry point
  9. pandapower/timeseries/run_time_series.py — Time series entry point
  10. pandapower/networks/power_system_test_cases.py — Built-in test networks
Retrieved from "http://wiki.gi7b.org/index.php?title=Pandapower_Architecture&oldid=329"