Record the measurement.
State the assumptions.

Coordinates and the measurement model

Use a fixed foundation reference. The origin is the left base edge; x increases right and y increases up. Positive horizontal displacement is right, positive vertical displacement is up, and positive rotation θ is clockwise. H is wall height. Top sensors measure at y = H; base sensors measure at y = 0. A sensor's x coordinate is measured from the left base edge.

Small-displacement, fixed-foundation-reference model: top horizontal = s + H*theta + r; base horizontal = s; base vertical at x = v0 - x*theta. Positive theta is clockwise, positive horizontal is right, positive vertical is up. r is remaining deformation, not an identified shear, bending or damage mechanism.

The recipe assumes small displacement and rotation, negligible movement of the reference frame, representative base translation and a base vertical-displacement pattern consistent with a rigid rotation and vertical translation. Local compression, foundation deformation, inconsistent attachment heights, sensor-axis misalignment and out-of-plane effects can violate those assumptions. A fitted rotation is an estimate under this recipe.

The remaining term r is top horizontal displacement after subtracting the supported sliding and rotation contributions. It is not automatically panel shear, bending, opening distortion, cracking or damage. Recording an opening on the diagram does not add a local opening model.

What the layout can identify

The planner constructs a matrix with one row per enabled instrument and checks its numerical rank. A top-horizontal row is [1, 1, 1, 0], a base-horizontal row [1, 0, 0, 0], and a base-vertical row [0, −x/H, 0, 1]. It reports identifiable parameter combinations as well as full rank. A complete decomposition normally needs top displacement, base sliding and vertical readings at distinct base locations. More sensors at one location do not create the missing geometric information.

Planning includes enabled instruments before they are mapped to data. Actual processing additionally requires available mapped readings at each time. Reusing the same acquisition column, including an alias of the same source file, cannot create independent measurements. A complete layout does not guarantee valid attachments, correct calibration or physical adequacy.

The solver uses equal-weight least squares with a singular-value decomposition. The reported condition number uses the stated q basis, so it describes this scaled measurement matrix. Closely spaced vertical instruments can strongly amplify noise. Review the geometry and sensitivity rather than interpreting rank alone as a quality certificate.

Calibration and raw source records

Imported CSV files are numeric tables with a unique header row. Blank measurements remain missing; entirely blank CSV records are ignored by the table parser while their original bytes remain in the source file. Invalid numerical text, inconsistent rows and unsupported file sizes are rejected; input rows are not silently sorted or deduplicated. The original bytes are retained alongside the recipe and SHA-256 fingerprint.

Gain is positive; polarity is +1 or −1. Displacement channels support mm and m, force supports N and kN, and time supports s and ms. Canonical outputs use mm, kN, s and radians. Check the acquisition system's calibration before selecting units or entering another gain; applying a calibration twice changes the result.

Confirm calibration and the model assumptions deliberately. Derived quantities remain unavailable until those declarations are recorded. The source citation, locator, permission statement, geometry, assignments and corrections travel with the project.

Clock alignment and missing data

Project time in seconds equals the source time converted to seconds plus its declared clock offset. Every configured source needs strictly increasing, finite time values. Choose one reference source to supply the output timeline. Offsets are user declarations; the Studio does not estimate clock drift or prove synchronization.

Exact alignment uses matching project times. Optional linear alignment uses the bracketing readings only within the declared maximum gap. It does not extrapolate outside a source, bridge blank endpoints or replace missing readings with zero. Interpolation changes the available information and is part of the saved recipe. Inspect the original clock scales and sampling intervals.

The workspace applies the recorded calibration and optional time interpolation. It does not silently smooth, filter, zero or resample the original file. Raw source bytes remain unchanged.

Uncertainty and signed percentages

Optional channel uncertainty is a one-standard uncertainty in the raw channel units. Supply a value that includes the relevant calibration and zero uncertainty. The recipe treats declared gains, geometry and clock alignment as fixed. Its propagated result assumes independent channel errors; shared-reference and systematic errors can violate that assumption.

Known channel uncertainties are propagated through the actual least-squares coefficients. A quantity affected by unknown input uncertainty has an unknown uncertainty. Uncertainty is withheld for interpolated readings because temporal covariance and interpolation-model error are not supplied. No confidence interval or validation certificate is inferred.

Component percentages are signed ratios to measured top displacement. They can be negative or exceed 100% when terms oppose each other. They are suppressed near zero top displacement, using the greater of your positive displacement floor and three times known top-displacement uncertainty. If top uncertainty is unknown, only the user floor is available; it does not establish a noise-based threshold.

Replay and observations

The common timeline follows actual calculated sample times. Plots and a selected-value table share that cursor. Any displayed wall movement is a labelled schematic of supported base translation and rotation; it does not reconstruct panel or crack deformation. Display magnification is a visual scale.

Attach video locally and record the project time corresponding to video time zero. Video time = project time − that anchor. This is a manual offset, without automatic clock-drift correction. Browser codec support can vary. Video bytes are never included in a saved project; reattach the file after reopening.

Observations are manual labels and notes at declared times, optionally linked to a selected photograph. A load drop alone does not establish cracking, failure mechanism or connection damage. Keep physical interpretations tied to observations and source limitations.

Save, reopen and export

Save the project ZIP to retain the exact CSV sources, recipe and selected raster photographs. Reopening verifies declared sizes and hashes and recalculates derived results. Supported project ZIPs use stored entries generated by this workspace. Hashes establish matching contents; files and recipes remain editable and are not signed proof of authenticity.

Export the full calculated timeline as CSV and a printable HTML report containing the recipe, assumptions, layout, summary and observations. Browser Print can produce a PDF. A report does not independently validate the measurements or calculation choices.

Limits: 4 CSV sources, 2 MiB per source, 20,000 data rows and 32 columns per source; 16 instruments and 100 observations. Up to 8 PNG, JPEG or WebP photographs, 3 MiB each and 12 MiB total, with a 20-megapixel decoding limit. Project ZIPs are limited to 24 MiB.

Research context and current scope

Deformation decomposition is established experimental practice. For context, Aloisio and co-authors discuss explicit assumptions when separating wall displacement contributions in Plastic deformation contributions of CLT and LTF shear walls (WCTE 2023). The Studio's limited base-kinematics recipe is not an implementation of that paper's full analysis or capacity model.

The existing seven Atlas curves do not contain the multi-channel instrumentation needed for these results. Import authorized channel data from your own work. Neither an empty dataset nor missing sensors establish a research finding.

Files are processed in your browser and are not automatically uploaded, published or backed up by this workspace. Use your existing LingoScholar account to enter, then save a project ZIP to keep the work.