tunnels and train length added by Roli15 · Pull Request #10 · dkouzoup/ms-eetc

Roli15 · 2026-05-25T18:17:47Z

No description provided.

# Conflicts: # train.py

dkouzoup

Thanks for the great work! I think this is a good point in time to add a small unit testing framework in python to ensure that all functionality works as expected and that old functionality was not affected. I suggest you start with something small (add a 'tests' directory and some first unit tests to test results with train length) and we can work incrementally from there.

dkouzoup · 2026-05-26T07:57:35Z

        # NOTE: good idea vel0 to be compatible with f0 (power-wise) to avoid nans at first iteration

-        vel0 = (60/3.6)**2
+        vel_avg = (self.points.index[-1]-self.points.index[0])/(terminalTime-initialTime)


I understand that this may be a smarter initialization but we should make sure that we do not break any of the paper simulation scripts. I would suggest once this MR is ready to run all the simulations generating the paper figures once again locally (it's fine if the results change slightly)

dkouzoup · 2026-05-26T09:46:25Z


+        if not self.crossSectionsOk():
+
+            raise ValueError("Issue with track cross sections!")


maybe "Issue with tunnel cross sections!"?

dkouzoup · 2026-05-26T09:50:33Z

+
+        self.tunnels.drop(columns=["Length [m]"], inplace=True)
+
+        self.crossSections = self.tunnels


this assignment and deletion seems unnecessary, why not calling it crossSections from the start?

removed reassignment

dkouzoup · 2026-05-26T09:52:44Z

+            if not any(abs((p - e)) < 0.1 for p in positions)
+        ]
+
+        openTrack_df = pd.DataFrame({"Length [m]": 0.0, "CrossSection [m^2]": float("inf")}, index=endOfTunnelPositions)


I would use a different name because openTrack is a known software tool and might be misleading at first glance

called it nonTunnelSections_df

dkouzoup · 2026-05-26T10:07:17Z

-        curvature = ca.MX.sym('curvature')
+        gradient = ca.MX.sym('gradient')                        # [-]     -> values between [0,0.2]
+        gradientLinearTerm = ca.MX.sym('gradientLinearTerm')    # [1/m]
+        curvature = ca.MX.sym('curvature')                      # [1/m]   -> values between [0,0.004]


Is there maybe a typo in some of the units here? I see that 1/m for gradient linear term becomes 1 after multiplying with position, but I would expect that the curvatureLinearTerm to have a different unit in that case

I corrected the units

dkouzoup · 2026-05-26T10:14:58Z


            tApprox = model.states[0]

+            epsVelSq = 0.0001


I see that structurally we need to change the code to compensate for the new state, but can you explain why we need this tolerance now?

I had the problem, that without the max operator IPOPT ran into negative values during probing. We can have a closer look at it together

Yes, let's please discuss this one. I just noticed that you introduced also fmax operator.

dkouzoup · 2026-05-26T11:08:35Z


+def computeTunnelFactor(cross_section, train):
+
+    if cross_section == 0:


I thought before you were using the convention that inf means open track? I would say inf or None are more suitable than 0 for no tunnel.

yes, you are right, should be inf

dkouzoup · 2026-05-26T11:15:00Z

+    num_of_remaining_points = numIntervals + 1 - len(requiredPoints)
+    m = 1 # number of points to oversample
+
+    while True:


interesting implementation, perhaps it could be extended to guarantee a minimum space interval?

dkouzoup · 2026-05-29T07:05:46Z

+            ],
+            [
+                2000.0,
+                Infinity,


Please use "infinity" (incl. the quotes) as defined in TTOBench

(unless there is a strong reason against it, in that case we'd need to update TTOBench instead)

dkouzoup · 2026-05-29T07:07:41Z

@@ -0,0 +1,96 @@
+{
+    "metadata": {
+        "id": "CH_Stadler_FLIRT_TPF",


just a detail: please use CH_Stadler_FLIRT_TPF also in filename (capitals) to be consistent with TTOBench and with file metadata

dkouzoup · 2026-05-29T07:12:03Z

+
+            plt.show()
+
+    def testAllIntegratorTypesWork(self):


probably this should not be in the trainLengthDependentTrackAttributes directory?

dkouzoup · 2026-05-29T07:13:29Z

+
+            plt.show()
+
+    def testAllIntegratorTypesWork(self):


why do we need this twice?

dkouzoup · 2026-05-29T07:14:29Z

+        Track with 20 permil increase from 1000 m to 2000 m and 20 permil
+        decrease from 3000 m to 4000 m.
+
+        Energy consumption should be roughly equal if computed using piecewise


I think the patern "solve the same OCP, roughly same consumption" can be improved in tests like this. We can discuss it.

Roli15 added 12 commits May 20, 2026 20:51

tunnel factor added

8a0efc0

speed limits train length adjusted

379d221

wip

c79639e

wip 2

9b6bc8c

fix tunnel parameter

518780b

Merge branch 'recovered-work'

2265ec6

# Conflicts: # train.py

gradients length dependent

3c32acd

fixed time approx

9776a58

curvature train length dependent

9b9c212

bug fixes

43b6e16

bug fix

0e20613

length dependent constant values adjusted

0089887

dkouzoup requested changes May 26, 2026

View reviewed changes

Roli15 added 8 commits May 26, 2026 13:50

tunnel resistance made to list

b1b1566

update tunnel resistance format

52fdc5e

test tunnel resistance added

6a0b452

test speed limit added

97c20ab

test gradients added

1a2a00b

test curvature added

8fe528e

implementation of pull request feedback

6aab544

bug fixes

32d3c76

dkouzoup reviewed May 29, 2026

View reviewed changes


		if not self.crossSectionsOk():

		raise ValueError("Issue with track cross sections!")


		self.tunnels.drop(columns=["Length [m]"], inplace=True)

		self.crossSections = self.tunnels


		def computeTunnelFactor(cross_section, train):

		if cross_section == 0:

Conversation

Roli15 commented May 25, 2026

Uh oh!

dkouzoup left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

dkouzoup May 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

dkouzoup May 29, 2026 •

edited

Loading