[Hayakawa, M.Takahashi, Sawada, Omae, Yamaguchi, R.Takahashi]

We continued the preparation. The clean booth was lowered with the chain blocks to access the roof. The bridge was set between the footholds. We removed the fastener part, which was fixed with double-sided tape and bolted Al plates.

https://gwdet.icrr.u-tokyo.ac.jp/~controls/slowmonitor/RT/

1. Yearly CSV Partitioning (2020–2026)

• Consolidation & Backfill: Recovered 2020–2025 archived data from backup and backfilled the missing 11-day gap (2026-01-01 to 01-12, 256 hours) from NDS2 (k1nds0:8088).
• Files Created: data_RT_2020.csv through data_RT_2026.csv, keeping data_RT.csv synced with 2026 for compatibility.
• Web Links: Added [2026] to [2020] download links to RT/index.html.
• Script Update: Updated plot_RT.py to auto-append to data_RT_{YYYY}.csv and load the previous year for continuous 100-day/1-year trends.

2. Interactive Plot Filtering

• Time Windows: Enabled toggling for all 6 generated spans (1 hour, 1 day, 1 week, 1 month, 100 days, 1 year).
  • Default: 1h, 1d, 1w, 1m ON; 100d, 1y OFF.
• Categories: Added toggles for BOOTH, FIELD, COOLER, and other (Default: all ON).
• Controls & UX:
  • Added [Select All] / [Clear] per row, and global [All ON], [All OFF], [Reset to Default].
  • Tied section header visibility directly to whether plots are displayed (eliminating orphan headings).
  • Added a status banner when either Time Window or Category is completely cleared.
  • Persisted user settings across page refreshes via localStorage.
• Language: Unified all buttons and labels to English.

3. Cache & Permissions

• Replaced cache-store with no-cache, no-store, must-revalidate and added .htaccess.
• Synchronized production (/users/PEM/SEM/, /mnt/slowmonitor/) and dev (/users/PEM/SEM_dev/).
• Set chmod 444 on RT/index.html.

. Glitch Filtering & Missing Data Handling

• Non-physical Glitch Removal:
  • Implemented an in-memory rolling-median filter (15-sample window, 10 hPa10\text{ hPa}10 hPa threshold, physical bounds [800,1150] hPa[800, 1150]\text{ hPa}[800,1150] hPa) in plot_AIR.py for all trend and summary plotting.
  • Successfully eliminated isolated packet dropouts and transient CS sensor drops (∼100 hPa\sim 100\text{ hPa}∼100 hPa downward glitches on Oct 19 and Oct 26) without altering raw CSV records.
• ATOTSU Station Outage Handling:
  • Handled persistent zero values from the ATOTSU weather station (0 hPa0\text{ hPa}0 hPa since mid-March 2026 due to station PC failure) by rendering a clear No Data (Station Down / PC Failure) placeholder on the 7-day trend plot while preserving valid historical data in the 1-year view.

2. Amplitude Spectral Density (ASD) Display Fix

• NaN Handling in Welch FFT:
  • Resolved the blank ASD spectrum issue by linearly interpolating isolated 1-minute dropouts on active channels and automatically skipping channels with >50%>50\%>50% missing data (ATOTSU).
  • Restored full physical air pressure spectra (2×10−6 Hz2\times 10^{-6}\text{ Hz}2×10−6 Hz to 8×10−3 Hz8\times 10^{-3}\text{ Hz}8×10−3 Hz) in AIR_ASD.png.
• Color & Legend Harmonization:
  • Synchronized colors and legend labels between the weekly summary (AIR_summary.png) and the ASD spectrum (AIR_ASD.png):
    • PRESSURE_MOZUMI: Blue
    • PRESSURE_CS: Orange
    • PRESSURE_EXC: Green
    • PRESSURE_EYC: Red
    • PRESSURE_ATOTSU: Gray (reserved for when active)

3. Yearly CSV Partitioning & Historical Consolidation

• Historical Data Consolidation:
  • Merged archived data from data_backup20260101 (2024–2025) and current 2026 data.
  • Backfilled the 2.5-month gap (2026-01-01 to 2026-03-15) from NDS2, establishing a complete and continuous record.
• Partitioned CSV Files:
  • Generated yearly files in /users/PEM/SEM/data/ and /mnt/slowmonitor/data/:
    • data_AIR_2024.csv (2024-05-15 to 2024-12-31)
    • data_AIR_2025.csv (2025-01-01 to 2025-12-31)
    • data_AIR_2026.csv (2026-01-01 to present)
    • data_AIR.csv (maintained as the latest year for backward compatibility)
• Automated Rollover & Multi-Year Trends:
  • Updated plot_AIR.py to automatically append new measurements to data_AIR_{YYYY}.csv.
  • Implemented multi-year data concatenation for 100-day and 1-year plots so spans crossing year boundaries remain seamless.
• Web Interface:
  • Updated index.html with direct download links for 2026, 2025, 2024, and Latest CSV files.

4. Browser Cache Management

• Replaced the caching header (Cache-Control: store) in index.html with explicit anti-cache directives (no-cache, no-store, must-revalidate).
• Added .htaccess and image version query parameters to ensure browser reloads immediately display up-to-date plots.
• Verified identical synchronization across /users/PEM/SEM/, /users/PEM/SEM_dev/, and /mnt/slowmonitor/.

Abstract:

I recovered IMC with ASC.
To keep the beam axis to the IFO, I changed MCE trans QPD1 and QPD2 setpoints as follows:
QPD1P: No change
QPD2P: No change
QPD1Y: From 0.003 to -0.043
QPD2Y: From 0.006 to -0.062

ASC for IMC can be engaged but both pitch and yaw of IP1 PZT was saturated, so PZT offload inside the PSL is necessary.

Detail:

After changing the beam polarization from P to S, IMC lock can be recovered without any issues.
So, I started the recovery of IMC alignment.

What we would like to achieve this time is to compensate for the change in cavity axis due to the MCE shift of ~10.26 mm by adjusting the MCI and MCO alignments to -3.7 and +3.7 urad, respectively (klog37708), while keeping the beam axis aligned to the IFO.
So, QPD signals should be changed as follows:
IMMT1T QPD: no change to keep the beam spot on IMMT1.
MCET QPD1: no change in pitch and some change in yaw.
MCET QPD2: no change in pitch and some change in yaw.

To check how much amount we need to change the yaw setpoints of MCET QPD1 and QPD2, I moved the MCI and MCO suspension by 3.7 urad and measured the signal changes in MCE QPD1 and QPD2.
Following tables are the results:

Our target is -3.7 urad for MCI and 3.7 urad for MCO, so QPD1 and QPD2 signals should be changed by -0.046 and -0.068, respectively.
Since the initial setpoints for QPD1Y and QPD2Y are 0.003 and 0.006, respectively, QPD setpoints should be -0.043 and -0.062.

After changing the QPD setpoints, I engaged the IMC ASC and all beam spots on QPDs can be controlled.
However, IP1 output is saturated and input beam axis cannot be aligned to the IMC axis, we need to offload the PZTs inside the PSL room.

In addition, MCI and MCO OpLevs are at the edge of linear range (or slightly off), so OpLev centering should be done at some point.
Figure 1 and 2 show the gigE camera image at IMC REFL and MCE trans, respectively for the reference.

PV:K1:VAC-PRESSURE_X_ITMX has been added to the MEDM screen “VAC_OVERVIEW”.

We removed the inserted HWP and restored the polarization to its original state.

1. Prediction Line Continuity Fix

• Issue: The MCF forecast line (dashed) showed an unphysical 2-order-of-magnitude vertical gap from the measured data (solid) at current time tlastt_{\text{last}}tlast​.
• Cause: Forecast baseline y0y_0y0​ used the regression intercept (s*0 + icpt) instead of the actual latest observation.
• Fix: Anchored y0y_0y0​ directly to the latest measured value (y0=log⁡10(Platest)y_0 = \log_{10}(P_{\text{latest}})y0​=log10​(Platest​)), guaranteeing continuous transitions.

2. Adaptive Window for Physical Regime Changes (TMP Turn-On)

• Issue: MCF had a sudden physical pressure plunge at ~08:50 JST when the turbo molecular pump (TMP) was started. A fixed 6-hour window included the pre-TMP roughing plateau and transition plunge, overestimating the pump-down slope (s=−0.58 dec/hs = -0.58\text{ dec/h}s=−0.58 dec/h vs. actual −0.15 dec/h-0.15\text{ dec/h}−0.15 dec/h).
• Fix: Implemented backward adaptive scanning from tlastt_{\text{last}}tlast​ using a normalized rate metric relative to 10−7 Pa10^{-7}\text{ Pa}10−7 Pa base vacuum: Metric=∣Δlog⁡10PΔt∣log⁡10(P/10−7 Pa)\text{Metric} = \frac{\left|\frac{\Delta \log_{10} P}{\Delta t}\right|}{\log_{10}(P / 10^{-7}\text{ Pa})}Metric=log10​(P/10−7 Pa)​ΔtΔlog10​P​​​
  • Avoids any singularity at P=1 PaP = 1\text{ Pa}P=1 Pa (log⁡101=0\log_{10} 1 = 0log10​1=0) while scaling appropriately across both atmospheric and ultra-high vacuum ranges.
  • Automatically isolates the post-TMP smooth regime (09:00 onwards, ~4 hours), yielding accurate physical slopes while retaining the full 6-hour window for steady channels.

3. Filtering of Non-Physical Communication Spikes (Glitch Rejection)

• Issue: Isolated 1-minute downward spikes (e.g. simultaneous drop by a factor of 40 across all channels at 2026-10-26 16:02 JST due to EPICS/NDS2 buffer packet dropouts) produced visual needles on plots and risked corrupting the latest observation.
• Fix: Implemented 5-point rolling log-median outlier rejection (clean_spikes and clean_spikes_tsd) with a 0.5 decade0.5\text{ decade}0.5 decade threshold.
  • Applied immediately before plotting (Plot_TSD, Plot_Trend, Vacuum_summary.png) and forecast calculation, while keeping raw historical records intact in data_VAC.csv.
  • Effectively removed all communication glitch needles across all plots without affecting real physical pump-down curves.

4. Synchronization and Verification

• Synchronized changes between /users/PEM/SEM/script/plot_VAC.py and /users/PEM/SEM_dev/script/plot_VAC.py.
• Re-executed the pipeline and verified clean, spike-free plots and continuous prediction curves in /mnt/slowmonitor/VAC/Vacuum_summary.png.

https://gwdet.icrr.u-tokyo.ac.jp/~controls/slowmonitor/Road41/

1. Problem & Root Causes

• Plot Failure: The legacy MLIT observation URL (rain84452105.html) was decommissioned (404), and tenki.jp forecast parsing crashed on decimal rainfall values (ValueError: invalid literal for int: '0.1').
• Data Gap: Historical observation tables were no longer available from the source, leaving a gap since 2026-10-22.

2. Actions Taken

• Forecast Parsing Fix (plot_Road41.py): Used pd.to_numeric(..., errors='coerce') to properly handle decimal precipitation.
• New Observation Source: Replaced the scraper with the lightweight MLIT mobile portal (fp116KishoInfoDetail_844500018_84.html) using certifi.where() for GlobalSign Root R46 SSL verification.
• Incremental Data Accumulation:
  • Implemented self-managed logging to append single-timestamp records into data_Road41.csv and web/Road41/data_Road41.csv.
  • Treated cumulative rainfall - as 0.0 mm when the hourly rate is 0 mm/h.
• Web UI & Cache Fix (index.html): Updated source links, added a direct CSV download link, and added cache-busting headers/JavaScript for Road41_rain.png.
• 10-Minute Automation (wget.sh): Integrated plot_Road41.py with CCTV image fetching. Updated crontab to 5,15,25,35,45,55 * * * * to match MLIT's 10-minute update cycle with a 5-minute buffer.
• Environment Sync: Synchronized all scripts and HTML files between /users/PEM/SEM/ and /users/PEM/SEM_dev/.

3. Current Status

• Verified automated pipeline: observations and CCTV images refresh cleanly every 10 minutes, appending new records to data_Road41.csv and generating updated plots.

https://gwdet.icrr.u-tokyo.ac.jp/~controls/slowmonitor/Earthquake/

1. Problem Summary

• The Earthquake monitoring page (https://gwdet.icrr.u-tokyo.ac.jp/~controls/slowmonitor/Earthquake/) had been frozen since January 3, 2026.
• The hourly cron job running plot_Earthquake.py continuously failed, preventing all plots (EarthquakeInfo.png, Earthquake_map.png, *SHINDO*.png) and earthquake tables from updating.
• Browsers continued displaying obsolete plots due to improper client-side caching settings.

2. Root Causes

• Corrupted CSV Record:
  • In /users/PEM/SEM/data/info_EQ.csv (lines 6491–6492), a historical earthquake entry from 2024-01-06 was split into two lines (新潟県上中,,,,,,, followed by .7,37.1,...).
  • In get_JMA(), pd.to_datetime(df2['JST']) attempted to parse 37.1 as a datetime string, raising ValueError: Given date string not likely a datetime. and crashing the script before any plot generation.
• Lack of Defensive Exception Handling:
  • Datetime parsing lacked error handling (errors='coerce').
  • HTTP requests fetching JMA XML feeds lacked timeout specifications, risking indefinite script hangs upon network drops.
• Improper Cache-Control Configuration:
  • web/Earthquake/index.html had <meta http-equiv="Cache-Control" content="store">, causing browsers to aggressively cache HTML and plot images.

3. Actions Taken

• Data File Repair:
  • Backed up info_EQ.csv and merged the split line back into its original valid format (...新潟県上中越沖,2.7,37.1...) in both data/info_EQ.csv and web/Earthquake/info_EQ.csv.
• Script Hardening (plot_Earthquake.py):
  • Used pd.to_datetime(..., errors='coerce') combined with dropna(subset=['JST']) so unexpected data corruptions do not halt the entire pipeline.
  • Added timeout=20 to urllib.request.urlopen and wrapped feed/item parsing in try-except blocks.
  • Added safe handling (.fillna(0).astype('int')) for M4+ event list filtering.
  • Synchronized the hardened script to /users/PEM/SEM_dev/script/plot_Earthquake.py.
• Browser Cache Prevention (web/Earthquake/index.html):
  • Updated meta headers to no-cache, no-store, must-revalidate and Expires: 0.
  • Injected a lightweight JavaScript cache-buster to append dynamic timestamps (?t=...) to plot image URLs on page load.
  • Synchronized updated HTML to SEM_dev.

4. Final Status

• plot_Earthquake.py executes cleanly with exit code 0.
• Latest earthquake feeds from JMA (up to October 28, 2026) and real-time KAGRA Shindo trends from NDS are successfully merged.
• All plots and summary files (EarthquakeInfo.png, Earthquake_map.png, info_EQ_over-M4.txt, ACC_PSL_SHINDO-*.png, SEIS_*_SHINDO-*.png) are regenerated and now automatically refresh in browsers.

I checked the recent HTCondor logs for Gauch on the DetChar cluster and found that the plotting job had been on hold since October 25.
The job was held because its CPU usage (11 CPUs) exceeded the requested number of CPUs (10 CPUs).
I modified the code to add diagnostic logs to identify which part of the multiprocessing operations causes the CPU usage to exceed the requested limit.

In addition, I removed the currently submitted HTCondor jobs to free up the DetChar cluster for Oshino-san's work. Once Oshino-san has finished the work on the cluster, I will resubmit the jobs.

Problems

• The VAC page stopped updating after 2025-01-31, when the legacy vacuum DAQ (vac_now.dat via K1VAC_PRESSURE_DAQ.py) stopped.
• chlist_VAC.csv still listed the legacy channels. TimeSeriesDict.fetch fails entirely if one channel is missing, so no data was fetched.
• The empty DataFrame crashed plot_VAC.py (IndexError) and truncated data_VAC.csv to 3 bytes.
• Plot_Trend raised ZeroDivisionError for gauges reading 0 Pa.
• The summary legend was cluttered with inactive channels.
• The forecast was a quadratic fit in log-log space. It extrapolated upward even for clearly monotonic pump-down data.
• Some changes in SEM/ were reverted at 16:13 (chlist_VAC.csv, pySEM.py, backfill_VAC.py), and the cause is unknown. SEM_dev and SEM were also both running SEM.sh from cron against the same log.

Actions

• Replaced chlist_VAC.csv with 25 active channels (CS 8, X 10, Y 7), matching VAC_OVERVIEW.adl. The old list is saved as .bak_old.
• pySEM.py changes:
  • get_TSD() falls back to per-channel fetch.
  • Plot_Trend is guarded against vmin <= 0.
• plot_VAC.py changes:
  • CSV writes are protected against an empty DataFrame.
  • The legend shows only channels with data in the last day, in 2 columns.
  • The forecast was rewritten as a damped log-linear extrapolation (details below).
• Restored the pre-2025 history from SEM_dev, then backfilled 2025-02-01 to 2026-10-18 with backfill_VAC.py. The 10-day recent window was processed first.
• Rewrote index.html with the new per-chamber gauge links (MCF, IFI, PR2, PR3, BS-SR2, GVs, X/Y arms).
• Re-synced chlist, pySEM.py and plot_VAC.py between SEM and SEM_dev after the revert. data_VAC.csv was restored from /mnt/slowmonitor/data/.
• Removed the SEM_dev SEM.sh cron line, leaving only the SEM hourly job at :07.
• The new forecast works as follows:
  • The slope of log10(P) is estimated over the last 6 h with Theil-Sen.
  • The slope decays with a 12 h time constant.
  • The result is clipped to 1e-8–101325 Pa.
  • Decreasing, increasing and flat trends were checked on synthetic data.

Final state

• data_VAC.csv is about 19 MB and 35,867 hourly rows, continuous from 2024 to Oct 2026.
• All subplots and the summary plot show current data.
• The MCF pump-down forecast decays monotonically and levels off, and increasing trends are extrapolated upward.
• The hourly cron updates the page.

Remaining items

• backfill_VAC.py is a 0-byte file, with no copy in SEM_dev. It is no longer needed because the backfill is complete.
• The forecast does not reproduce oscillations. It gives a flat mean-level line, and a seasonal (daily) component can be added later if needed.
• The cause of the 16:13 revert is not identified.

[Hayakawa, M.Takahashi, Sawada, Omae, Yamaguchi, R.Takahashi]

We prepared to repair the clean booth roof. We covered the top chamber and the TMP with the Seiden sheet. The footholds were set by the chamber. The chain blocks were tested.

1. Issues Identified

• Service Outage: The Lightning Monitor page had stopped updating since April 2026.
• Corrupted & Monolithic Data:
  • The primary data file (data_Blitzortung.csv) had accumulated over 2.4 GB of data without yearly partitioning.
  • Broken rows (e.g., incomplete comma-separated lines) caused pandas parsing crashes during routine processing.
• Redundant & Fragmented Web UI:
  • The stroke histogram was displayed twice on the webpage, and plots for lightning strokes and KAGRA underground magnetometers (p2p) were rendered as separate, unaligned figures.
  • No direct download links existed for yearly event archives.
  • Overlapping colors between the histogram and magnetometer lines created visual ambiguity.
• Update Synchronization Lag:
  • 24-hour location maps were updated every 10 minutes, but the stroke histogram was generated only hourly, creating temporary discrepancies where recent strikes appeared on the map but were absent from the histogram.

2. Actions Taken

• Data Cleanup & Yearly Archiving:
  • Sanitized corrupted records from existing Blitzortung logs.
  • Split the monolithic dataset into per-year archive files (data_Blitzortung_2020.csv through 2026.csv) and created matching near-range lists (EventList_300km_YYYY.csv).
  • Updated scripts to automatically append to yearly files and seamlessly handle cross-year queries (past 8-day rolling window) across New Year transitions.
• Unified Plot & Visual Redesign:
  • Merged the Blitzortung stroke histogram and KAGRA underground magnetometer (p2p) into a single 2-panel plot sharing the same x-axis (Lightning_Mag_ptp.png).
  • Removed the redundant title between axes and embedded subplot titles inside the legend boxes.
  • Re-colored the lightning distance bins to Red (<100 km), Yellow (100–300 km), and Gray (300–1000 km) to prevent confusion with the magnetometer traces (Blue: CS, Orange: X-end, Green: Y-end).
• Pipeline Synchronization & Update Timestamps:
  • Implemented magnetometer p2p data caching (mag_p2p_cache.csv) and timestamp tracking files (mag_updated.txt, blitz_updated.txt).
  • Updated Read_Blitzortung.py to regenerate the unified combined plot every 10 minutes along with the 24-hour maps, eliminating time lag between map and histogram.
  • Displayed the exact JST last-update time for each dataset directly inside their respective legend titles (Update: YYYY-MM-DD HH:MM JST).
• Webpage (index.html) Cleanup:
  • Removed duplicate standalone histogram images.
  • Inserted automated navigation links for yearly event CSV archives ([2026] [2025] ... [2020]).

3. Final Status

• Both scheduled jobs (Read_Blitzortung.sh every 10 min and SEM.sh hourly) execute successfully with zero errors.
• The web page (Lightning & Thunder Monitor) displays synchronized 24-hour strike maps, yearly CSV links, and the unified 7-day Lightning & Magnetic Field plot with real-time JST update timestamps.

I aligned IMC before MC chamber vacuum evacuation (fig1).

I measured the TFs of the MCe suspension after connecting the bellows. The length-to-yaw TF was consistent with the reference (Fig. 5). Compared with the TFs before connecting the bellows, the pitch-to-yaw coupling became larger than the pitch-to-length coupling (Fig. 6). The behavior of the suspension looks healthy.

I measured the TFs of the MCe suspension after removing the clean bag. The coupling to the yaw mode was large (Fig. 1-4). The resonant frequency of the pitch mode was a little lower than the reference (Fig. 6). The behavior of the suspension looks healthy.

I remeasured the TFs of the MCi suspension after aligning the mirror to the Oplev QPD. The suspension behavior looks healthy.