incremental-well-validation.html

<!doctype html>
<html lang="en">
<head>
<meta charset="utf-8"><meta name="viewport" content="width=device-width, initial-scale=1">
<title>Incoming well signals · Complete numerical review</title>
<style>
:root{font-family:system-ui,sans-serif;color:#162d38;background:#f2f5f4;line-height:1.5}*{box-sizing:border-box}body{margin:0}main{max-width:1120px;margin:auto;padding:28px 24px 64px}h1{font-size:clamp(2rem,5vw,3rem);line-height:1.15;max-width:760px;margin:.4em 0}h2{font-size:1.35rem;margin-top:0}h3{font-size:1rem}a{color:#085a65}p{max-width:850px}section{margin:24px 0;padding:22px;background:#fff;border:1px solid #cfdbd7;border-radius:8px}.eyebrow{font-size:.8rem;font-weight:700;letter-spacing:.08em;text-transform:uppercase;color:#42626d}.notice{border-left:5px solid #ad641d;background:#fff5e7;padding:14px 18px;margin:20px 0}.controls{display:flex;gap:18px;flex-wrap:wrap}label{display:flex;flex:1;min-width:180px;flex-direction:column;gap:6px;font-weight:600}select{font:inherit;padding:10px;width:100%;background:white;border:1px solid #647d85;border-radius:5px}.flow{display:grid;grid-template-columns:repeat(4,1fr);gap:12px}.flow>div{padding:14px;border:1px solid #cad8d2;background:#eef4f1;border-radius:4px}.flow strong{display:block}.flow span{font-size:.9rem}.facts{display:grid;grid-template-columns:repeat(3,1fr);gap:18px}.fact strong{display:block;font-size:1.6rem}.detail{font-size:.92rem;color:#42606a}.fail{color:#912f20;font-weight:650}.pass{color:#185a48}.scroll{overflow-x:auto}table{border-collapse:collapse;width:100%;font-size:.9rem}caption{text-align:left;margin:10px 0;color:#42606a}th,td{padding:10px 9px;border-bottom:1px solid #d8e0dc;text-align:left;vertical-align:top}thead{background:#f2f6f4}code{overflow-wrap:anywhere;font-size:.85rem}.links{display:flex;gap:20px;flex-wrap:wrap}.selected{background:#fff3df}[hidden]{display:none!important}.error{color:#912f20}footer{padding-top:20px;font-size:.9rem;color:#42606a}@media(max-width:650px){main{padding:20px 16px 40px}section{padding:16px}.facts{grid-template-columns:1fr}.flow{grid-template-columns:1fr 1fr}.controls{flex-direction:column}label{min-width:0}th,td{padding:8px 7px}h1{font-size:2rem}}
</style>
<script src="incremental-well-validation.js" defer></script>
</head>
<body><main>
<a href="tcn-stream-validation.html">← Earlier complete score replay</a>
<div class="eyebrow">Well surveillance · Saved research results</div>
<h1>Does the incoming signal path preserve the original model inputs?</h1>
<p>This review follows every incoming row through a bounded history, model features, one unchanged checkpoint and its frozen alarm. Choose a checkpoint and recording to inspect the actual results.</p>
<div class="notice"><strong>Industrial deployment is unqualified.</strong> This explorer displays saved validation evidence. It receives no telemetry and runs no model. Physical calibration, trusted sources, useful fresh detection, target-device tests and site acceptance remain open.</div>
<p id="load-state" role="status">Loading the complete saved review…</p>
<div id="audit" hidden>
<section aria-labelledby="scope-title"><h2 id="scope-title">Complete scope, preserved failures</h2>
<div class="facts"><div class="fact"><strong id="source-rows"></strong>original source rows<div class="detail">33 recordings / eight filename groups</div></div><div class="fact"><strong id="timing-rows"></strong>accepted updates timed<div class="detail">Two separate passes over the same data</div></div><div class="fact"><strong id="execution-state"></strong>provider execution<div class="detail" id="conformance-state"></div></div></div>
<p id="original-failure" class="detail"></p><p class="detail">The second attempt fixes only Arrow conversion of null sensor cells. Data, weights, scalers, thresholds and strict tolerances remain unchanged. Both attempts survive; reused validation is not new independent confirmation.</p>
</section>
<section aria-labelledby="path-title"><h2 id="path-title">What the measured path includes</h2><div class="flow"><div><strong>1. Packet</strong><span>Declared sequence, time, channels and numeric availability</span></div><div><strong>2. History</strong><span>512 past samples; a 64-sample tail; no future values</span></div><div><strong>3. Model</strong><span>Eight identical copies of only the current feature input</span></div><div><strong>4. Alarm</strong><span>Unchanged threshold and streaming alarm state</span></div></div>
<p class="detail">Eight copies add computation and establish no new observations or model promotion. The measured packet-to-alarm timer excludes file decoding and diagnostic storage. Arrival times and units are retrospective declarations; this is not a physical acquisition or device deadline test.</p>
</section>
<section aria-labelledby="selection-title"><h2 id="selection-title">Inspect one complete recording</h2><div class="controls"><label>Checkpoint<select id="checkpoint"></select></label><label>Recording<select id="recording"></select></label></div><p id="selection-note"></p>
<div class="facts"><div class="fact"><strong id="score-failures"></strong>scores outside original tolerance<div class="detail" id="score-detail"></div></div><div class="fact"><strong id="alarm-changes"></strong>changed alarm decisions<div class="detail" id="alarm-detail"></div></div><div class="fact"><strong id="pipeline-median"></strong>eligible packet-to-alarm median<div class="detail" id="pipeline-tail"></div></div></div>
<p class="detail" id="missing-now"></p>
<div class="scroll"><table><caption>Every feature comparison stays visible. Bit differences and failures at the original tolerance are different measurements.</caption><thead><tr><th>Feature</th><th>Compared values</th><th>Outside tolerance</th><th>Different bits</th><th>Maximum finite difference</th></tr></thead><tbody id="feature-rows"></tbody></table></div>
<p class="detail" id="discrete-changes"></p><p class="detail" id="alarm-usefulness"></p>
</section>
<section aria-labelledby="group-title"><h2 id="group-title">All eight groups</h2><p class="detail">These are published filename groups. Independent physical-well identities are not authenticated. No group or failed recording is omitted.</p><div class="scroll"><table><thead><tr><th>Group</th><th>Historically eligible rows</th><th>Feature failures</th><th>Score failures</th><th>Alarm changes</th><th>Eligible with current missing channels</th></tr></thead><tbody id="group-rows"></tbody></table></div></section>
<section aria-labelledby="evidence-title"><h2 id="evidence-title">Traceable evidence</h2><p id="runtime"></p><p class="detail">The 28,672-byte ring holds numeric history only. Process-lifetime peak memory includes the reference and diagnostic harness; neither is a deployed edge memory requirement. Raw timers, first calls and outliers remain in the original output.</p><div class="links"><a id="report-link">Measured report</a><a id="output-link">Original recovery output</a><a id="rebuild-link">Saved-array replay guide</a></div><p>Immutable scientific revision: <code id="artifact-revision"></code></p><p>Recovery output SHA256: <code id="output-sha"></code></p><p class="detail">No fits, retuning or reserved-well predictions occurred. Both original and later numerical failures remain. No model is promoted, and zero alarm changes do not establish a useful detector.</p></section>
</div><footer>Original adaptation and author review, with AI assistance. Petrobras 3W data retain CC BY 4.0 attribution; authored methods are MIT. No independent physical or human signoff is claimed.</footer>
</main></body></html>