Class NetworkValidationBenchmarks

java.lang.Object
neqsim.process.equipment.network.NetworkValidationBenchmarks
public class NetworkValidationBenchmarks extends Object
Validation benchmark cases for pipeline network solvers.

Provides analytically solvable or published benchmark cases for verifying the Hardy Cross and Newton-Raphson solvers against known solutions. Each benchmark returns a result object with computed values, expected values, and pass/fail status.

Benchmark Cases

  • Single Pipe (Darcy-Weisbach): Analytical ΔP from Swamee-Jain / Darcy-Weisbach equation
  • Two Parallel Pipes: Known flow split from equal/unequal diameter pipes
  • Triangle Loop: Classic 3-pipe loop with Hardy Cross analytical solution
  • Two-Loop Network: Cross (1936) textbook example with 5 pipes and 2 loops
  • Mass Balance: Verifies conservation of mass at all junction nodes
  • Pressure Monotonicity: Verifies pressure drops along flow direction

Reference: Cross, H. (1936). "Analysis of Flow in Networks of Conduits or Conductors". Bulletin 286, University of Illinois Engineering Experiment Station.

Version:
1.0
Author:
Even Solbraa
See Also:

  • Field Details

    • logger

      private static final org.apache.logging.log4j.Logger logger

  • Constructor Details

    • NetworkValidationBenchmarks

      public NetworkValidationBenchmarks()

  • Method Details

    • createBenchmarkGas

      private static SystemInterface createBenchmarkGas()
      Default gas for benchmarks: methane-dominated natural gas at 50 bar, 25 C.
      Returns:
      configured gas system

    • runSinglePipeBenchmark

      public static NetworkValidationBenchmarks.BenchmarkResult runSinglePipeBenchmark()
      Benchmark 1: Single pipe Darcy-Weisbach pressure drop.

      Verifies: For a single straight pipe of known L, D, roughness, with a given mass flow, the network solver produces the same pressure drop as the Darcy-Weisbach equation with the Swamee-Jain friction factor.

      Analytical solution: 

      f = 0.25 / [log10(e/(3.7D) + 5.74/Re^0.9)]^2   (Swamee-Jain)
dP = f * (L/D) * (rho * v^2 / 2)
      Returns:
      benchmark result

    • runParallelPipeBenchmark

      public static NetworkValidationBenchmarks.BenchmarkResult runParallelPipeBenchmark()
      Benchmark 2: Two parallel pipes with known flow split.

      For two parallel pipes with the same L and roughness but different diameters D1 and D2, the flow split at steady state satisfies equal pressure drop across both paths: 

      dP1(Q1) = dP2(Q2) and Q1 + Q2 = Q_total

      For turbulent Darcy-Weisbach: Q1/Q2 ≈ (D1/D2)^(5/2) (approximate for equal friction factors).

      Returns:
      benchmark result

    • runTriangleMassBalance

      public static NetworkValidationBenchmarks.BenchmarkResult runTriangleMassBalance()
      Benchmark 3: Triangle loop mass balance.

      A triangle network (A-B-C-A) with one source at A and two sinks at B and C. Verifies that mass is conserved at all junction nodes and that the Hardy Cross solver correctly distributes flow.

      Returns:
      benchmark result

    • runSolverCrossVerification

      public static NetworkValidationBenchmarks.BenchmarkResult runSolverCrossVerification()
      Benchmark 4: Hardy Cross vs Newton-Raphson agreement.

      Solves the same two-loop network with both methods and verifies that results agree within tolerance. This is a cross-verification benchmark — both solvers should converge to the same physical solution.

      Returns:
      benchmark result

    • runPressureMonotonicity

      public static NetworkValidationBenchmarks.BenchmarkResult runPressureMonotonicity()
      Benchmark 5: Pressure monotonicity along flow direction.

      In a purely pipe-based network with no compressors, pressure must decrease along the flow direction. This benchmark verifies that the solver produces physically consistent pressure profiles.

      Returns:
      benchmark result

    • runSparseVsDenseBenchmark

      public static NetworkValidationBenchmarks.BenchmarkResult runSparseVsDenseBenchmark()
      Benchmark 6: Sparse vs Dense solver agreement for large networks.

      Constructs a 10x10 grid network (100 nodes, ~200 pipes) and verifies that the sparse CSC solver and dense Gaussian elimination produce identical results. Reports timing for both.

      Returns:
      benchmark result

    • runAllBenchmarks

      public static List<NetworkValidationBenchmarks.BenchmarkResult> runAllBenchmarks()
      Run all benchmarks and return aggregate results.
      Returns:
      list of all benchmark results

    • buildTwoLoopNetwork

      private static LoopedPipeNetwork buildTwoLoopNetwork(SystemInterface gas, String prefix)
      Build a standard two-loop network for benchmarking.
      Parameters:
      gas - fluid template
      prefix - name prefix
      Returns:
      configured network