US 10247599

I'll search for authoritative information on this patent number and check for any litigation or CAFC docket activity.

Let me look for the claim language and any litigation activity specifically tied to this number.

Let me check the litigation/docket databases and confirm the full claim set.

US Patent 10,247,599 — Analyst Summary

Important scope caveat: I could not directly query the USPTO Patent Center or the CAFC docket system from my tools — I have no litigation-database access, and my searches returned only patent-document sources (Google Patents, Justia, uspto.report). So the bibliographic data below is well-grounded, but the "litigation status" finding is a negative search result, not a verified docket check. Treat it as low-confidence.

Bibliographic data

Note: the application is a national-stage entry (15/328,308 series, "328" indicating a §371 filing) of a PCT application, which is why the filing date (2015-08-18) is nearly a year after the Japanese priority date.

Abstract (as published)

"A rotating machine abnormality detection device and a rotating machine abnormality detection method according to the present invention are designed to detect an abnormality in a rotating machine including a rotary member rotatable about a predetermined axis, and detect an initial abnormality in the rotating machine based on first measurement data obtained by measuring a vibration in an ultrasonic wave band caused by the rotary member. A rotating machine according to the present invention includes the rotating machine abnormality detection device."

Plain-language overview of the independent claims

Based on the claim text reproduced in the record (Google Patents and the Justia copy of the pre-grant publication at https://patents.justia.com/patent/20170219420), the patent appears to have three independent claims — one apparatus claim, one system/machine claim, and one method claim:

1. Detection device (independent claim 1). A device for detecting abnormality in a rotating machine having a rotary member that rotates about an axis. It comprises (a) an ultrasonic wave measuring section (e.g., an AE/acoustic-emission sensor) that measures vibration in the ultrasonic band caused by the rotary member, and (b) an abnormality detecting section that detects an initial abnormality based on "first measurement data" from that ultrasonic sensor. The core inventive hook is the use of the ultrasonic band — rather than the audible band — so that steady machine characteristic vibrations don't mask an early-stage fault.

2. Rotating machine (independent claim). A rotating machine (motor, generator, compressor, pump, etc.) that incorporates any one of the above-described abnormality detection devices.

3. Detection method (independent claim). A method with two steps: an ultrasonic wave measuring step (measuring ultrasonic-band vibration caused by the rotary member) and an abnormality detecting step (detecting an initial abnormality from the resulting first measurement data).

Dependent-claim subject matter (plain language)

The dependent claims build a two-stage architecture:

• Stage 1 — detect early, in ultrasound. Compute a first frequency spectrum of the ultrasonic first measurement data; compute a first feature amount — specifically the integrated value over a predetermined frequency range; compare it to a threshold Th1; declare an initial abnormality when the integrated value meets or exceeds Th1. The claims also note that the "initial abnormality" may be tied to the time of occurrence.
• Stage 2 — classify, in audible band. Add a vibration measuring section (e.g., accelerometer) for the audible band; after the ultrasound-triggered detection, determine abnormality type and/or degree based on second measurement data. This is done by comparing a second frequency spectrum computed from data before detection (normal pattern) with a third frequency spectrum computed from data after detection (abnormal pattern) — i.e., a before/after spectral comparison.
• Specific diagnostic taxonomy. Where the rotary member is a pair of intermeshing sub-rotary members (male/female screw rotors) in a casing, the claims distinguish a first abnormality (rotor-to-rotor contact) from a second abnormality (rotor-to-casing contact) — corresponding respectively to spectral intensity changes at the rotor rotational frequencies and their harmonic relationship (CT1×b = CT2×a).

Litigation / CAFC 2026 docket status

• No litigation or Federal Circuit docket entry was found linking US 10,247,599 to any proceeding. Multiple targeted searches ("10247599" combined with "litigation," "District Court," "Federal Circuit," "Unified Patents," "Docket Alarm," "Kobe Steel") returned no case records. The only CAFC-related hit was an unrelated appeal docketed July 8, 2026, Contour IP Holding LLC v. GoPro, Inc., Fed. Cir. No. 2026-2012 — which involves different patents.
• Do not read this as confirmation of non-litigation. The absence of a hit in my search tooling is weak evidence, and I could not query PACER, Docket Alarm, or the CAFC's own docket. If a litigation search matters to your analysis, this should be re-run against a dedicated litigation database.

Identifier disambiguation (per the strict-interpretation rule)

The string "10247599" appears in several unrelated contexts, which should not be conflated with the US patent:

• DE 10247599 A — an unrelated German published application cited in an EPO search report for EP 1554388 (plant gene expression).
• A Brazilian corporate registration number "10247599 SCHAAN PARTICIPACOES SOCIETARIAS LTDA" (JUCISRS filing, Feb. 2024).
• An entry in a UK Gazette notice list.

Residual uncertainties

1. Exact claim count and numbering. The record I retrieved reproduces the claims in prose form (via Justia and the Google Patents full text) but I did not obtain a clean numbered claim list from the face of the granted patent. I identified three independents; a fourth independent claim cannot be ruled out.
2. Assignee currency. The record shows Kobe Steel, Ltd. as current assignee with reassignment recorded in 2017. I have no evidence of any later assignment, but I also cannot confirm a current chain of title.
3. The "expires 2036-01-28" figure is Google Patents' computed adjusted-expiration estimate, not an official USPTO determination.

I'll search for litigation involving US patent 10247599 across patent litigation databases.

Initial searches returned irrelevant results. Let me search the litigation-specific databases directly.

Let me try more targeted searches on litigation-specific portals and docket databases.

Litigation Search Results: US Patent 10,247,599

Bottom Line

No litigation involving U.S. Patent No. 10,247,599 has been identified. My searches across patent-litigation resources returned no complaints, docket entries, PTAB proceedings, or appellate cases in which this patent was asserted or challenged.

Before reporting this, let me be precise about what the patent is, so we can be confident the search was directed at the right asset:

This is a Japanese-origin (Kobe Steel) industrial machine-diagnostics patent covering ultrasonic/AE-based detection of incipient abnormalities in rotating machinery (e.g., screw compressors). It is a single-patent family member with a U.S. pre-grant publication (US 2017/0219420 A1) and corresponding foreign filings — not a patent-family with a large U.S. assertion history typical of a PAE campaign.

Searches Performed

I searched for the literal number "10247599" and "10,247,599" combined with litigation-, docket-, and infringement-related terms (including the Unified Patents litigation portal, CourtListener, and general case-law/docket sources). The relevant and irrelevant hit patterns were:

• Only false positives appeared, none involving this patent. Examples:
  • A Brazilian corporate-registry filing listing "10247599 SCHAAN PARTICIPACOES SOCIETARIAS LTDA" (unrelated company registration number).
  • A German prior-art reference "DE 10247599 A" (a 2002 German application, unrelated technology).
  • A UK Gazette notice number "10247599" (unrelated).
  • References to different U.S. patents with overlapping last digits (e.g., the '475 patent in AACI v. ..., US 4,237,454, US 7,842,595) — these are not US 10,247,599 and must not be conflated.
• The Google Patents page for US 10,247,599 shows a legal-status box (active, expiring ~2036-01-28) and an assignment record (recorded 2017-01-23 to Kabushiki Kaisha Kobe Seiko Sho), but no "Litigation" section — Google Patents populates that section when it detects asserted patents in its litigation dataset, so its absence is consistent with no known U.S. assertions.

Cases Found

None. There are therefore no plaintiffs, defendants, jurisdictions, case numbers, filing dates, or outcomes to report.

Important Caveats (as a matter of candor, not comfort)

1. This is a negative search result, not a certification. Absence in the sources reachable via the available search tooling does not conclusively establish that no litigation exists. The most reliable databases for a definitive answer — Docket Navigator, Lex Machina, RPX Insight, and the PACER/Unified Patents "caselist" — are largely paywalled or not fully indexed by general web search, which limited what I could retrieve.
2. Possible non-U.S. or non-assertion activity not captured. Because the patent is held by an operating company (Kobe Steel) rather than an NPE, its more likely involvements would be (a) defensive — e.g., an IPR/PGR petition filed against it, or a DJ action; or (b) foreign (e.g., Japanese) enforcement. My searches did not surface any PTAB petition bearing this patent number either. Note that the patent's filing date (2015) post-dates the AIA, so it is IPR-eligible.
3. A related-family or continuation patent could be the asserted asset. It is worth checking whether Kobe Steel holds continuation/sibling U.S. patents in this family and whether those have been asserted, since a litigation search keyed strictly to "10247599" would miss a sibling.

Recommended Next Steps for a Definitive Answer

To convert this negative result into a reliable clearance-style conclusion, run the following targeted checks:

• Docket Navigator — party search on "Kobe Steel" as patent plaintiff/defendant, and a "patent number" query set to exactly 10,247,599.
• Lex Machina / RPX Insight — same exact-number query.
• USPTO Patent Trial and Appeal Board (PTAB) end-to-end / PTAB Center — search for the patent number in IPR, PGR, and Derivation proceedings.
• USPTO Patent Center "Litigation" / Global Dossier and the Unified Patents Portal patent page for US-10247599-B2.
• CourtListener/RECAP full-text docket search for "10,247,599" and "10247599."

If you have access to any of the paywalled databases above, I can help interpret the output or draft a formal no-litigation representation based on their results.

Disclosure of uncertainty: I could not retrieve any primary litigation record for this patent, and general web search returned only unrelated-number false positives. I am explicitly flagging that I do not have high-confidence, database-verified confirmation of a complete absence of litigation — only that no litigation is known from the sources searched.

I'll verify the canonical "no proceedings" default against public sources before drafting.

No hits so far — the '599 references are a different patent (8,489,599, Xerox). Let me run two more targeted checks before concluding.

One final check for any very recent filing or a defensive-aggregator (Unified Patents) challenge before I conclude.

Proceedings overview

Zero. The USPTO Open Data Portal structured block for US 10,247,599 returns no AIA trial proceedings — no IPRs, no PGRs, no CBMs, whether active, terminated, settled, or denied. My independent web checks (Google Patents, the patent's own PDF front matter, PTAB document aggregators, and litigation/PTAB coverage) surfaced nothing contradicting that, so the ODP result stands as the operative fact: all claims of the '599 patent are UNTESTED at the PTAB.

Bottom-line defensive posture for a defendant: this is the opposite of a hardened patent and also the opposite of a gutted one. Nothing has been canceled, so you cannot win by pointing to a PTAB disposition; but nothing has been litigated at the Board, so there is no § 315(e)(2) estoppel wall and the entire §§ 102/103 prior-art field is still open to you. Practically: a defendant's first question here isn't "which claims survived," it's "why has nobody bothered to challenge this yet" — which usually means the patent has not been asserted against a party with the incentive and budget to file.

⚠️ Disambiguation warning — do not let a search engine fool you. Searching "the '599 patent" + "IPR" returns a flood of PTAB and Federal Circuit material that has nothing to do with this patent. The heavily-litigated '599 is U.S. Patent 8,489,599 (Xerox / Palo Alto Research Center, "Context And Activity-Driven Content Delivery And Interaction"), which was the subject of IPR2021-00987, IPR2021-01294, and IPR2021-01458 (Snap, Twitter/X, and Facebook as petitioners), consolidated and decided by a consolidated Final Written Decision on 2023-03-24 holding claims 1–25 unpatentable under §§ 102/103, then appealed by Xerox to the Federal Circuit (CourtListener recap: https://storage.courtlistener.com/recap/gov.uscourts.cafc.20311/gov.uscourts.cafc.20311.26.0_1.pdf). None of that involves Kobe Steel or US 10,247,599. Any AI-generated summary or keyword search that conflates them is wrong.

No proceedings to report

Because the canonical list is empty, there are no proceeding entries to render in the requested per-proceeding format (no petitioner/patent-owner captions, no panels, no institution decisions, no FWDs, no settlements, no appeals). I will not manufacture proceeding numbers, judge names, or dispositions to fill the template.

For the record, the patent's own bibliographic facts (from the authoritative full text provided):

Strategic summary

Canceled vs. sustained vs. untested. There is no PTAB narrowing of any kind. Every claim stands as issued and is UNTESTED — not "sustained," which is a meaningful distinction: a claim that survived an IPR carries a presumption-enhancing litigation narrative, whereas an unchallenged claim has simply never been stress-tested. (I note I do not have a verified formal claim listing in the record provided to me, only the claim-style text reproduced in the summary-of-invention portion — an ultrasonic-wave measuring section + abnormality detecting section for detecting initial abnormality; the frequency-spectrum/integrated-value/threshold limitations; the audible-band vibration measuring section with pre-/post-detection comparison; the pair-of-sub-rotary-members-plus-casing limitations covering contact-between-rotors and rotor-to-casing abnormalities; a rotating-machine claim; and a method claim. I am not assigning claim numbers to those, because the claim set text was not in the source and guessing would be fabrication. Pull the granted claims from the USPTO PatentCenter record before relying on any numbering.)

Estoppel landscape. Section 315(e)(2) estoppel is a nullity here — estoppel only attaches to a petitioner that reached a final written decision. No petitioner, no FWD, no estoppel. That is unambiguously good news for a would-be challenger: there is no procedural bar, no prior panel has construed a single term, and no petitioner has locked up an obviousness combination. Conversely, there is also no IPR record to borrow from — no claim constructions, no expert findings, no admissions by the patent owner. Any invalidity case has to be built from scratch.

Pattern signals. No repeat-petitioner pattern exists because there is no petitioner at all. No defensive aggregator (Unified Patents or similar) appears anywhere in the chain for this patent. No PTAB appeal history, because there is no PTAB history. I also found no district court or ITC assertion of US 10,247,599 in public sources — the "defendant" framing in this assignment is therefore somewhat hypothetical. A caveat on confidence: district court dockets and recently-filed complaints are imperfectly indexed by web search, so a PACER / Docket Navigator check is warranted before treating "never asserted" as established. Sectorally, this is an industrial condition-monitoring patent (screw-compressor rotor-contact detection via AE sensors, FFT band-integral thresholding vs. audible-band spectrum comparison) owned by a large Japanese industrial manufacturer — the profile suggests cross-licensing or competitor-facing enforcement rather than a volume-assertion campaign. The family is broad: siblings include US 10,054,481; US 10,401,329; US 10,422,774; US 10,352,760; US 10,385,853; US 10,620,090; US 11,060,902; and US 11,143,632. I did not exhaustively verify PTAB activity on those siblings, so treat "no IPRs on the family" as unconfirmed.

Recommended next steps

If you are a defendant and want to invalidate. There is no FWD to link and no disposition to quote — say that plainly to the client rather than implying a prior win. The absence of PTAB activity is itself the signal. Two concrete art leads lie on the face of the record:

• The patent's own front matter lists a European search report (EPO, dated 2018-04-18, re EP 15839521.0) citing Yongbo Shao et al., "Development of acoustic emission and vibration monitoring instrument for rotating machinery," Proc. SPIE, 2004-10-18, pp. 317–321 — i.e., AE-plus-vibration monitoring of rotating machinery, the exact two-sensor architecture of the challenged claims.
• Sibling US 10,054,481 B2 (same family, same assignee's AE-diagnosis line) cites directly on-point rotor-contact art: JP H08-166330, JP H09-133577, and JP 2010-234403, plus the patent's own cited JP 2001-074616 (Patent Literature 2) and JP 2012-208045 (Patent Literature 1). Sources: https://patentimages.storage.googleapis.com/8e/af/2d/7986dea108c85a/US10422774.pdf and https://patentimages.storage.googleapis.com/3c/0c/21/552fba826724a6/US10247599.pdf
• Note the "initial abnormality via ultrasonic band" theory is the stated point of novelty over JP 2001-074616, which used wavelet-transform vibration analysis in the audible band. That admission in the specification is a useful obviousness hook (why would AE-based contact detection in the ultrasonic band have been non-obvious to a skilled artisan in 2014?).

Timing and deadlines. With no active proceeding, there is no institution deadline, no oral hearing date, and no statutory 1-year FWD clock to report. Two clocks that do matter: (1) the § 315(b) one-year bar from service of a complaint — if you have been served, that window is running and is the hard constraint on filing; and (2) the patent's adjusted expiration of 2036-01-28, meaning roughly a decade of remaining life, so a validity challenge is worth the spend if the patent is commercially material.

Verify before filing. Before any petition, confirm on USPTO PatentCenter / PTAB E2E (https://ptacts.uspto.gov/ptacts/) that no proceeding has been filed since the ODP ingest date, and confirm the exact granted claim set. My searches were consistent with the ODP block, but ODP ingest lags and a very recently filed petition may not appear.

I'll research the assignment chain for this patent. Let me start with targeted searches.

Assignment-chain reconstruction — US 10,247,599 B2

Tooling caveat (read first). I could not reach assignmentcenter.uspto.gov or assignment.uspto.gov from my toolset. Despite a large number of targeted queries (patent number, application number 15/328,308, title, inventor names, assignee names), I retrieved no reel/frame image, no assignment cover sheet, and no abstract-of-title data for this patent. Everything below is built from Google Patents legal-events data, USPTO/Google-published front-page data, EPO register data for the sibling family, and third-party patent databases. I am not going to invent reel/frame numbers. Treat the reel/frame fields as unverified; if you need them, pull the Abstract of Title from the Assignment Center search page directly: https://assignment.uspto.gov/patent/index.html#/patent/search/result?q=[10247599](/patent/10247599)

Inventors

Two inventor-site clusters appear (Kobe research labs and Takasago Works), which is normal for a machinery-diagnostics invention that spans controls R&D and rotating-equipment engineering.

Departure pattern — not present / no adverse signal. The strongest available evidence runs the opposite way. A continuation-family application, WO 2018/066267 (published 2018-04-19, priority 2016-10-05, app. 2017/030772), names Takahashi Eiji, Araki Kaname, Takaki Shugo plus Hayashi Masato — all Kobe Steel. In other words, three of the four inventors were still filing rotating-machine abnormality applications for Kobe Steel roughly two years after the 2014-09-12 priority date, and Takaki is separately credited with 12 granted US patents at Kobe Steel. There is no evidence of the team leaving the assignee within 12 months of filing. No fire-sale precursor signal here.

Original assignee

Kabushiki Kaisha Kobe Seiko Sho (Kobe Steel, Ltd.), 2-4, Wakinohama-Kaigandori 2-chome, Chuo-ku, Kobe-shi, Hyogo 651-8585, Japan (the older 3-18 / 10-26 Wakinohama-cho 1/2-chome address appears on older records).

• Primary line of business: integrated steel and metal products manufacturer (steel, aluminum/copper alloys, welding consumables, titanium), plus industrial machinery including compressors — the "Kobelco" brand. Founded 1905; listed on the Tokyo Stock Exchange.
• Did it ship a product embodying the claims? Yes, in the ordinary-course sense. The patent is a condition-monitoring / abnormality-diagnosis device for its own rotary compressors (screw compressors with male/female rotors, per FIGS. 2–3). Kobe Steel is both the machinery manufacturer and the intended user of the diagnostic. This is a manufacturing-process / product-reliability patent for a compressor maker, not a licensing asset.
• Current status: operating and publicly traded. Pertinent context: Kobe Steel was hit by a major product-data falsification scandal in October 2017 (falsified inspection data on aluminum and copper products), which triggered a global customer audit. Notably, that event did not produce a patent portfolio liquidation — Kobe Steel continued filing in this family (2016-priority filings; EP 3 206 003 granted 2019) and the patent remains in Kobe Steel's name.
• SEC angle: Kobe Steel is a Japanese issuer; it does not file US 10-K/8-K with the SEC. I found no SEC filing evidencing a transfer of this patent. Its annual securities reports (有価証券報告書) are filed with JPX/EDINET, not SEC EDGAR.

Assignment timeline

I could identify one recorded transfer of rights, plus a chain of prosecution/legal events. I could not retrieve the reel/frame, the correspondent of record, or the exact recording date from a primary source.

• Executed 2017-01-23 (recording date shown by Google Patents legal events; execution date itself not verifiable from my sources) — Reel/Frame NOT VERIFIED (could not retrieve)
  • Conveyance: Assignment of assignors' interest (inventors → company)
  • Assignor: ARAKI, Kaname; MATSUMOTO, Yasuharu; TAKAHASHI, Eiji; Takaki, Shugo
  • Assignee: KABUSHIKI KAISHA KOBE SEIKO SHO (KOBE STEEL, LTD.)
  • Correspondent: not retrieved. Related but distinct: the US associate firm handling this same Kobe Steel rotating-machine family in prosecution was Studebaker & Brackett PC (recorded as agent of record on sibling US 10,401,329 B2 = US 15/327,895, filed 2015-08-18). I flag this only as a pointer for where to look for the assignment cover sheet — I have no evidence it is the assignment correspondent, and a single appearance would not be a finding anyway.
  • Context: ordinary-course in-house assignment — the standard obligation-to-assign instrument by which salaried inventors vest title in their employer. Not a sale, not a reorg.

Everything else on the record is a legal event, not an ownership transfer:

• 2015-08-18 — US national-stage application filed (§371 entry of the PCT; hence the ~11-month lag behind the 2014-09-12 JP priority).
• 2017-08-03 — Pre-grant publication US 2017/0219420 A1.
• 2019-04-02 — Patent granted.
• Google Patents lists Current Assignee = Kobe Steel, Ltd. with legal status Active, adjusted expiration 2036-01-28 (a Google-computed estimate, not a USPTO determination).

No post-issuance assignment (no re-assignment, security agreement, merger, name change, or license) appears in any source I can reach. That absence is itself the finding: the patent has never left Kobe Steel according to the public record.

Timeline diagram

timeline
    title Ownership of US 10247599
    2014 : JP priority filed by Kobe Steel inventors
    2015 : US national stage filed
    2017 : Inventors assign rights to Kobe Steel
         : Pre-grant publication
    2019 : Patent granted to Kobe Steel
    2026 : No later assignment found

NPE / troll-pattern signals

1. Shell-entity transfer — not present. The only recorded transfer runs into the operating company (inventors → Kobe Steel, 2017-01-23). Current assignee per Google Patents is still Kobe Steel, Ltd. No "IP/Licensing/Holdings/Ventures" suffix, no registered-agent address, no single-purpose LLC anywhere in the chain. There is simply no transfer out to examine.
2. Known asserter in the chain — not present. Neither the inventors nor Kobe Steel, Ltd. matches any public NPE list (Acacia, Marathon, IV, IPNav, Wi-LAN, Conversant/Mosaid, Vringo, Pendrell, Round Rock, MPHJ, Lumen View, Spangenberg entities). Kobe Steel is a 120-year-old steel manufacturer.
3. Repeat correspondent across the chain — unclear (unverifiable). The chain has only one link, so "recurrence" cannot be tested. I could not obtain the recorded correspondent for the 2017-01-23 assignment. The only firm I can associate with this family is Studebaker & Brackett PC (prosecution counsel of record on sibling US 10,401,329), and a single firm doing ordinary prosecution work for a Japanese manufacturer is not a finding under your recurrence standard.
4. Cascading transfers — not present. One assignment in ~12 years of file history; nothing resembling a <24-month LLC daisy-chain.
5. Pre-litigation transfer — not present. No infringement suit naming US 10,247,599 was found, and the only assignment on record is from 2017 — long before any hypothesized assertion and in the wrong direction (to, not from, the manufacturer).
6. Bankruptcy fire-sale — not present. Kobe Steel has not filed Chapter 7/11 and is not in insolvency proceedings. The 2017 data-falsification scandal is a reputational/regulatory event, not a bankruptcy, and no patent sale followed it in this family.
7. Privateering — not present. No transfer to a third-party assertion vehicle; no SEC or press coverage of Kobe Steel funding litigation against competitors via an NPE.
8. Defensive aggregator (anti-NPE) — not present. The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. Kobe Steel is the end of the chain, and it is an operating company, not a defensive patent fund.

Verdict

Insufficient data.

The entire ownership history consists of a single obligation-to-assign instrument (inventors → Kobe Steel, 2017-01-23, reel/frame not retrievable through my tools), with no subsequent transfer recorded in any source I can reach and Kobe Steel, Ltd. shown as current assignee with the patent Active. There is no shell entity, no known asserter, no cascading transfer, no bankruptcy sale, and no privateering arrangement — every NPE indicator is either "not present" or "unverifiable because the chain has only one link." On the record available, this reads as a normal captive industrial patent held by its operating-company manufacturer, not as an assertion asset. The "insufficient data" call reflects my inability to confirm reel/frame and correspondent from the Assignment Center, not any hint of an NPE pattern.

Verify at: https://assignment.uspto.gov/patent/index.html#/patent/search/result?q=10247599 — then cross-check the Abstract of Title and the recorded correspondent against the Google Patents legal-events entry (2017-01-23 reassignment, assignors Araki / Matsumoto / Takahashi / Takaki).

I'll research the citation records for US 10,247,599, focusing on the references actually cited against it (front-page patent citations and the two Japanese Patent Literature references named in the specification).

I found the front-page "References Cited" list. Let me now retrieve the complete citation list (including any U.S. patent documents above the JP entries) and details on each reference.

Preliminary note on method and evidence quality

I ran targeted searches against Google Patents, the machine-readable front page of the granted patent, J-GLOBAL, FreePatentsOnline, and the Unified Patents portal. I do not have USPTO Patent Center / PAIR or a litigation database in my toolset, so I cannot literally "query the USPTO database." What I can do — and did — is retrieve the "References Cited" block printed on the face of US 10,247,599 from the patent PDF, which is the USPTO-examined citation record reproduced verbatim.

Source for the citation block: http://patentimages.storage.googleapis.com/3f/0c/21/552fba826724a6/US10247599.pdf (front page, "References Cited").

Two important caveats before the substance:

1. The list I retrieved may be truncated. The snippet I captured begins at the first foreign document ("JP H08-043193 A"). It is common for the "References Cited" block to open with a "U.S. PATENT DOCUMENTS" sub-table above the foreign list, and I did not capture any U.S. patent documents. I cannot rule out that one or more U.S. patents are cited on the face of this patent. Do not treat the list below as exhaustive.
2. I did not capture the examiner's category tags (the * / "X" / "A" markers that flag which references were cited as anticipating vs. merely background). The front page of a U.S. patent does not use X/A categories; a "cited by examiner" asterisk marks examiner-added references. So my § 102 characterizations below are my own analysis, not a recitation of examiner findings.

Identifier disambiguation (strict-interpretation rule)

Per the operating rules, I am treating every alphanumeric string literally. The following are NOT US 10,247,599 and must not be conflated with it:

None of these are the subject patent, and none are citations of it.

The citation record of US 10,247,599

A. Patent documents cited (foreign)

B. Other publications cited

Items 7 and 8 are not prior art — they are the search reports generated during prosecution of this very family. They are listed as "other publications" per USPTO formality.

Per-reference analysis

Claim framework used below. From the granted text, the claim set is: (i) independent device claim (ultrasonic-band measuring section + abnormality detecting section detecting an initial abnormality from first measurement data); (ii) dependents adding the first frequency spectrum / integrated value over a frequency range vs. threshold Th1, and the "time of occurrence" limitation; (iii) dependents adding a vibration measuring section (audible band) and determination of abnormality type and/or degree; (iv) dependents reciting the pre-detection second spectrum vs. post-detection third spectrum comparison, with the second/third "feature amount" of the same type; (v) dependents reciting the pair of intermeshing sub-rotary members + casing, distinguishing a first abnormality (rotor–rotor contact) from a second abnormality (rotor–casing contact); (vi) an independent rotating machine claim; and (vii) an independent method claim. I was unable to obtain a clean numbered claim list from the face of the granted patent in the prior pass; the numbers are therefore described by class, not by numeral. Flagging this as an open item rather than guessing.

Statutory framing. Effective filing date = priority date 2014-09-12 (JP 2014-185873). The application was filed 2015-08-18, so the AIA § 102 regime governs. Every JP document above published before 2014-09-12, so each is available at minimum as a § 102(a)(1) printed publication.

1. JP H08-043193 A (Feb. 1996)

• Full citation: Japanese Unexamined Patent Application Publication No. HEI 8-043193 A (1996-02-16).
• Status: § 102(a)(1) art (published >18 yrs before the effective filing date).
• Description: I could not verify the subject matter of this document with the tools available. Do not treat this line as a description of its disclosure.
• § 102 potential: Undetermined. Because it is nearly 19 years pre-priority, it is chronologically eligible to anticipate any claim; whether it does cannot be assessed without its text. This is a gap that must be closed by pulling the JP full text (J-PlatPat) before any validity opinion.

2. JP H08-166330 A (Jun. 1996)

• Full citation: Japanese Unexamined Patent Application Publication No. HEI 8-166330 A (1996-06-25).
• Status: § 102(a)(1) art.
• Description: Unverified — no substantive content retrieved.
• § 102 potential: Undetermined; same gap as item 1.

3. JP 2001-074616 A (Mar. 2001) — this is "Patent Literature 2" of the patent

• Full citation (verified via J-GLOBAL): 特開2001-074616, "回転機の異常診断装置" (Rotating-machine abnormality diagnosis device). Application No. 特願平11-251905; filed 1999-09-06; published 2001-03-23. Agent: 大岩 増雄. Source: https://jglobal.jst.go.jp/en/detail?JGLOBAL_ID=200903081439337325
• Description (verified from the published claim/abstract): Detects the vibration of a rotating machine, applies a continuous wavelet transform, adds the transformed signal along the frequency axis, and decides abnormal/not-abnormal from the addition result. The abstract describes detecting a peak in a fault signal synchronized with the rotation period, and using the random (non-rotationally-synchronous) character of noise to suppress it. The patent's own BACKGROUND characterizes it as detecting bearing damage.
• § 102 potential:
  • Independent device claim (i): Does not anticipate. The reference operates on detected vibration (the patent treats this as the audible/structural-vibration regime), whereas the claim requires a vibration in the ultrasonic wave band. The patent expressly distinguishes this reference on exactly that ground ("the frequency band that can be recognized as a vibration is a band where a steady vibration such as a so-called characteristic vibration of the device overlaps"). Absent ultrasonic-band measurement, element (i)'s "ultrasonic wave measuring section" is missing → no § 102 anticipation.
  • Dependent claims on a frequency-spectrum feature amount (ii): Close, but not anticipatory. "Continuous wavelet transform + summation along the frequency axis" is a time-frequency analogue of "compute a frequency spectrum; integrate over a frequency range." But it is a wavelet transform with a frequency-axis sum, not an FFT-spectrum integrated value compared to a threshold Th1. This is § 103 material at most, and the patent's own specification distinguishes it (it cannot detect an initial abnormality because of characteristic-vibration overlap).
  • Independent method claim (vii): Not anticipated, for the same ultrasonic-band reason.

4. JP 2002-181038 A (Jun. 2002)

• Full citation: Japanese Unexamined Patent Application Publication No. 2002-181038 A (2002-06-26).
• Status: § 102(a)(1) art.
• Description: Unverified — no substantive content retrieved.
• § 102 potential: Undetermined.

5. JP 2010-234403 A (Oct. 2010)

• Full citation: Japanese Unexamined Patent Application Publication No. 2010-234403 A (2010-10-21).
• Status: § 102(a)(1) art (published ~4 years pre-priority).
• Description: Unverified — no substantive content retrieved.
• § 102 potential: Undetermined. Given its recency this is chronologically the most likely of the unverified group to be a genuinely pertinent § 102/§ 103 reference. Priority to verify.

6. JP 2012-208045 A (Oct. 2012) — this is "Patent Literature 1" of the patent

• Full citation: Japanese Unexamined Patent Application Publication No. 2012-208045 A (2012-10-25).
• Status: § 102(a)(1) art (published ~2 years pre-priority).
• Description: Per the patent's own BACKGROUND, this is the listening-rod (listening bar) inspection reference: a metal rod is placed against the machine and the other end against the inspector's ear; the inspector listens to the vibration noise and judges the presence/absence of an abnormal noise. It is a human-auditory, manual inspection method.
• § 102 potential:
  • Independent device claim (i): Does not anticipate. No "ultrasonic wave measuring section" and no automated "abnormality detecting section." It is manual and audible.
  • Dependent claims (ii)–(v): Not anticipated — no frequency spectrum, no threshold comparison, no second/third spectrum comparison, no automated abnormality-type taxonomy.
  • Independent method claim (vii): Not anticipated — the method steps ("ultrasonic wave measuring step" + "abnormality detecting step") are absent.
  • Role in prosecution: Background only. It is the reference the patent positions itself against (subjectivity/expertise dependence of the human inspector).

7. Yongbo Shao et al., Proceedings of SPIE, vol. 2620, pp. 317–321 — the most § 102-relevant item in the list

• Full citation as printed: Yongbo Shao et al., "Development of acoustic emission and vibration monitoring instrument for rotating machinery," Proceedings of SPIE, Oct. 18, 2004, pp. 317–321, vol. 2620.
• Status: § 102(a)(1) art (printed publication, well before 2014-09-12).
• Description: Unverified in full text. The title itself — "acoustic emission and vibration monitoring instrument for rotating machinery" — discloses the combination of (a) AE (acoustic-emission, i.e. ultrasonic-band) monitoring and (b) vibration monitoring applied to rotating machinery. That is precisely the two-sensor architecture that the patent's dependent claims add to the ultrasonic core.
• § 102 potential:
  • Independent device claim (i): Potentially the single most dangerous § 102 reference on this list, if the paper discloses detecting an abnormality from the AE signal. The claim's core is "ultrasonic wave measuring section for measuring a vibration in an ultrasonic wave band caused by the rotary member" + "abnormality detecting section for detecting an initial abnormality … based on first measurement data." An AE-monitoring instrument for rotating machinery plausibly meets the first element; whether it meets the "initial abnormality detecting section" element turns on the paper's processing disclosure, which I have not read.
  • Dependent claim adding the audible-band vibration measuring section: also plausibly met at the apparatus level (AE + vibration monitoring on the same machine).
  • Claims (ii), (iv), (v) (FFT spectrum + integrated value vs. Th1; pre/post spectrum comparison; rotor–rotor vs. rotor–casing taxonomy): almost certainly not anticipated — those are specific processing/diagnostic limitations unlikely to be in a 5-page SPIE instrument paper.
  • Recommendation: Obtain the actual paper (SPIE Vol. 2620, DOI/ISBN lookup) before relying on this list. I will not assert anticipation on a title alone.
  • Citation-accuracy flag: the printed data are internally suspicious. SPIE Vol. 2620 corresponds to a 1995 proceedings, not 2004, and the page range 317–321 is short for an instrument-development paper. Either the date, the volume, or both may be mis-printed on the patent face. This does not affect its status as prior art (either date precedes 2014-09-12), but it does affect your ability to retrieve it.

References I encountered that are NOT on the citation list but are worth flagging

Kobe Steel sibling applications (same priority date 2014-09-12)

The record shows a family of co-pending Kobe Steel filings with the identical 2014-09-12 priority date, including:

• US 10,288,475 B2 / US 2017/0184445 A1 — "Rotating machine abnormality detection device, rotating machine abnormality detection method, and rotating machine" (uses multiple sensors + phase difference to specify contact position; cites JP HEI 5-231361, JP HEI 9-133577, and JP 2001-74616 = the same PL2).
• US 10,401,329 B2 / US 2017/0212085 A1 (App. 15/327,895) — "…device, method and system…" (sampling + real-time FFT display).
• US 10,054,481 B2 — cites JP 2001-074616 and describes a rotor–rotor vs. rotor–casing contact distinguishing method (JP HEI 9-133577) using demodulated AE.

These siblings are co-owned, same-priority documents. Under AIA § 102(b)(2)(C) they are commonly-owned and would be excepted from § 102(a)(2) as to this patent — but they are highly relevant § 103/obviousness-double-patenting context because they show the applicant's own parallel development of the rotor-contact taxonomy. They are not in the "References Cited" block.

Unconfirmed possible citation: US 4,237,454

A Unified Patents portal result (https://portal.unifiedpatents.com/patents/patent/US-4237454-A) surfaced US 4,237,454 "System for Monitoring Bearings and Other Rotating Equipment" on a page also showing the title "Rotating Machine Abnormality Detection Device and Method and Rotating Machine." The relationship is ambiguous from the snippet — it may be a citation in the family, or merely a "similar documents" listing. I cannot confirm it is a citation of US 10,247,599. If it is, it is classic § 102(a)(1) art (1980) on AE-based bearing monitoring and deserves a full read.

Bottom line

Two contradictions/limitations I want to flag explicitly against the earlier summary:

1. The earlier summary stated the patent "appears to have three independent claims." That remains my reading, but it is still unverified against a numbered claim list — and my attempt to capture the face-of-patent claim set in this pass did not resolve it.
2. The earlier summary's litigation finding (no docket entries) stands unchanged and is still a negative search result, not a docket verification.

Three concrete follow-ups needed to make this a defensible prior-art opinion: (a) pull JP H08-043193, JP H08-166330, JP 2002-181038 and JP 2010-234403 full texts from J-PlatPat; (b) obtain the actual Shao SPIE paper and reconcile the volume/date discrepancy; (c) retrieve the un-truncated "References Cited" block (including any U.S. PATENT DOCUMENTS section) from Patent Center.

I'll ground this obviousness analysis in the prior art of record and in the field of acoustic-emission machine diagnostics. Let me run several targeted searches.

Excellent — the searches surfaced both the references of record (listed on the face of the '599 patent) and highly relevant art in the same field. Let me pull the remaining face-of-patent citations.

Obviousness Analysis — U.S. Patent No. 10,247,599 under 35 U.S.C. § 103

Building on the earlier "Patent summary" and "Litigation summary" sections; those bibliographic findings (priority 2014-09-12, filed 2015-08-18, Kobe Steel, three independents: device / rotating machine / method) are treated as given and are not repeated here.

0. Analytical framework and a threshold point

The '599 patent has a 2014-09-12 priority date, so the AIA (§ 102/103 as amended) governs. All art discussed below must therefore have been publicly available or "effectively filed" before 2014-09-12. I have flagged date-qualification issues where they exist.

Threshold point that shapes everything else: the "Prior Art" material for this patent is unusually self-defeating. The '599 specification itself frames the invention as a solution to a problem the applicant admits already existed in the art:

"…the frequency band that can be recognized as a vibration is a band where a steady vibration such as a so-called characteristic vibration of the device overlaps. This makes it difficult to detect an initial abnormality." (background, discussing PTL2)

That is an express, on-the-record statement of the problem — a classic "obvious to try" posture. Independently, the EPO's own search report on the family member EP 3 193 155 A1 flags JP 2002-181038 A (Mitsubishi Heavy Industries, 26.06.2002) with the "Y" category (relevant in combination) against claims 1–3, 9–10, and separately against claims 4–8. A "Y" citation means a European examiner already concluded that this art renders the claim set non-inventive when combined with another reference. That is objective corroboration of the § 103 case I develop below.

1. Analogous art — the field is saturated

Every reference below is in the same field of endeavor (machine condition monitoring) and reasonably pertinent to the problem the '599 claims address (detecting and classifying incipient rotor/bearing contact faults). No Wands/In re Bigio analogy problem exists.

Caveat: I could not retrieve the full text of JP H08-043193 A, JP H08-166330 A, or JP 2010-234403 A; descriptions of the H08 references are inferred from (a) the '599 and US 10,054,481 face citations and (b) the background of US 10,054,481, which summarizes the rotor-contact-via-rotational-frequency method. Flagged as inferred, not verified.

2. Claim 1 (and the machine/method independents)

Claim 1 requires only: (a) an ultrasonic wave measuring section measuring vibration in the ultrasonic band caused by the rotary member; and (b) an abnormality detecting section detecting an initial abnormality from the resulting first measurement data.

Ground 1A — JP 2002-181038 A alone (or EP 0209862; or US 4,790,190). Each discloses an AE/ultrasonic transducer mounted on rotating machinery, converting ultrasonic-band elastic waves into a signal, plus processing circuitry that outputs an abnormality determination. That maps onto (a) and (b) element-for-element.

Ground 1B — JP 2002-181038 A in view of JP 2001-074616 A (PTL2). PTL2 supplies the "abnormality detecting section" logic in generic form (detect vibration → transform → decide abnormal/not). Substituting an AE transducer for PTL2's accelerometer is a simple substitution of one known sensor type for another, producing the predictable result of sensitivity in a band where, as PTL2's own art field recognized, characteristic-vibration interference is absent. KSR Int'l v. Teleflex, 550 U.S. 398, 416 (2007).

The "initial abnormality" limitation is not a saving distinction. Three independent reasons:

1. It is a statement of result, not of structure or step. Claim 1 recites what is detected, not how. Where a claim recites a result and the prior art discloses the structure/process capable of producing that result, the claim is obvious; an inventor cannot claim the recognition of a result. In re Kollar; In re Montgomery.
2. The art already expressly taught incipient detection. US 5,072,611 states that ultrasonic AE monitoring "will provide earlier detection than standard thermal monitoring approaches," and that "…the acoustic signal will rise by about 10 to 12 dB as incipient bearing failure arises" — i.e., the ultrasonic band was used for initial fault detection thirty years before the priority date.
3. PTL2's admitted failure is the motivation, not a bar. The '599 specification concedes PTL2 cannot detect initial abnormalities because the audible band is masked. A reference that identifies a problem and points toward its cause motivates the very change claimed. That the change is to a band whose incipient-fault sensitivity was long known makes the combination obvious as a matter of law.

Method claim. The two-step method (measure ultrasonic vibration; detect initial abnormality) is performed by any of the above systems; process claims covering the operation of a known apparatus are obvious for the same reasons.

Machine claim. Appending "a rotating machine comprising [the device]" adds only the host machine (motor/generator/compressor/pump — all named in the '599 spec as conventional). In re Rose / obvious-design-choice.

3. Dependent claims — ultrasonic frequency spectrum, integrated value, threshold Th1

These claims add: compute a first frequency spectrum (FFT) of the ultrasonic data; compute a first feature amount = integrated value over a predetermined frequency range; compare to threshold Th1; declare initial abnormality when ≥ Th1; optionally treat the crossing time as the time of occurrence.

Ground 2 — Ground 1 + JP 3325448 B2 + JP 2001-074616 A.

• JP 3325448 B2 discloses precisely the thresholding architecture on AE data: compute the effective value of the AE detector output, derive a threshold for that value, compare, and count exceedances to judge the bearing. Substituting an integrated spectral energy over a band for an RMS time-domain value as the "feature amount" is an elementary signal-processing design choice — both are measures of signal energy, and JP 2002-098585 already teaches the spectral route.
• JP 2001-074616 A (PTL2) supplies frequency-domain reduction and threshold-based decision-making ("adds transform signals along the frequency axis… determines whether there is an abnormality based on an addition result").
• FFT itself was, by 2014, the most routine spectral tool in the field. JP 2002-098585 uses FFT on wavelet coefficients; US 2016/0055737 uses frequency-transformed inputs. No claim to non-obviousness can rest on using FFT.
• Time-of-occurrence: a natural and inescapable consequence of a threshold-crossing test on chronologically sampled data (the '599 spec itself derives the time of occurrence simply from "the time point at which the integrated value becomes equal to or greater than the threshold value").

Motivation: threshold-on-energy is the canonical AE condition-monitoring technique (JP 3325448; US 5,072,611's "predetermined level of increase… as an upper limit reference"). Reasonable expectation of success: high; this is measurement calibration, not a new principle.

4. Dependent claims — audible-band vibration sensor; abnormality type and/or degree after detection

These add a vibration measuring section (accelerometer) for the audible band, and require that after detecting the initial (ultrasonic) abnormality, the system determine abnormality type and/or degree from second measurement data.

Ground 3 — Ground 1 + JP 2001-074616 A (PTL2) and/or US 2016/0055737 A1 + WO 2012/088707 A1.

This is the weakest link in the patent's chain, because the combination of an AE sensor plus an accelerometer is not merely suggested — it is literally disclosed as a single sensor set:

• US 2016/0055737 A1 (Boken; prov. 2013-04-08): a sensor set "comprising an acoustic emission sensor, a magnetic pickup sensor, an accelerometer…"; the accelerometer "may also be referred to as a vibration sensor and is configured to provide the system with a vibration signal," while the AE sensor "may also be referred to as an ultrasonic sensor." The reference then compares an RMS statistic to a threshold against a normal-condition baseline, with the stated purpose of detecting problems "before catastrophic failures may occur." That is Ground 1 + the audible branch + the threshold/baseline mechanism in one document.
• WO 2012/088707 A1 likewise pairs a vibration sensor with an acoustic sensor and sets normal ranges against which each signal aspect is compared.
• JP 2005-062154 A expressly claims a sensor selected from "vibration sensor, ultrasonic sensor, AE sensor" — treating them as interchangeable/selectable alternatives, which is per se a teaching of substitutability.

Motivation to combine (the strongest available, and it is strong):

1. Complementary sensitivity is a matter of common engineering knowledge. AE/ultrasound is highly sensitive to incipient, contact-type, impulsive micro-events but (as the literature notes) of limited range and relatively insensitive to global mechanical faults; accelerometer-based vibration excels at developed faults (and at imbalance/misalignment). A POSITA pairing them to get early detection plus diagnostic discrimination is doing exactly what the art taught.
2. Explicit source-to-target directionality. JP 2002-098585 and US 2019/0049413 establish the paradigm "AE catches the early stage; the vibration method becomes informative as degradation develops" — the same two-stage architecture the '599 claims recite. (US 2019/0049413 is post-dated and is cited here only as evidence of the state of the art, not as § 102/103 prior art.)
3. The claimed sequencing is a routine engineering choice with a recognized benefit. "Determine type/degree after the ultrasonic trigger" merely avoids running the classifier continuously; the '599 spec itself frames the sole stated advantage as that it "can reduce redundant operational steps." Event-triggered analysis is a well-known design pattern; KSR ("a court must ask whether the improvement is more than the predictable use of prior art elements according to their established functions").

5. Dependent claims — pre-detection ("normal pattern") vs. post-detection ("abnormal pattern") spectra

These add: compute a second feature amount (second frequency spectrum) from audible-band data obtained before detection; a third feature amount of the same type from data after detection; and determine type/degree by comparing the two spectra.

Ground 4 — Ground 3 + JP 2002-098585 A + US 2016/0055737 A1.

• JP 2002-098585 A discloses the identical analytical move: comparing the frequency characteristic against a "normal-state pattern" (正常時パターン) and evaluating a degree of increase at a monitored frequency. That is the "normal pattern vs. abnormal pattern" comparison, verbatim in substance.
• US 2016/0055737 A1 discloses capturing a baseline RMS established "when the machine is known to be operating normally," storing it, and comparing future captures against it.
• The ring-buffer / FIFO retention of pre-trigger data is a routine data-structure choice (the '599 spec concedes it is merely "preferred"), and buffering is the universal way to obtain a pre-event record.

Motivation: before/after (or normal/current) signature comparison is the foundational paradigm of condition monitoring; nothing more than applying a known technique to the data already present.

"Degree increases as correlation value decreases" — correlation-based similarity scoring between spectra was elementary. Design choice.

6. Dependent claims — intermeshing sub-rotor pair; first abnormality (rotor↔rotor) vs. second abnormality (rotor↔casing)

This is the most technically specific subject matter in the patent: distinguishing rotor-rotor contact from rotor-casing contact via spectral intensity at rotational-frequency-related orders (CT1×b = CT2×a).

Ground 5 — Ground 3 + JP H08-166330 A and/or JP 2002-181038 A, optionally + JP 2005-062154 A.

This is where the record is most damaging to the '599 patent, because a sibling Kobe Steel patent, US 10,054,481 B2, describes this exact diagnostic as background art:

"…a diagnosing method for a compressor in which an AE wave generated by a contact between a male rotor and a female rotor… and rotational signals of the rotors are detected simultaneously… performing a frequency analysis and arithmetic mean processing to the output, and comparing the result with a determination reference on the basis of the rotational signals to perform diagnosis about a contact between the rotors…"

"Also, there has been proposed an irregular contact detecting method… demodulating an elastic wave signal output from the rotating machine, and determining that the rotors are in contact with each other if the demodulated elastic wave signal includes at least one of a rotational frequency component of the first rotor and a rotational frequency component of the second rotor at a predetermined or higher level, determining that a rotor is in contact with the casing if the demodulated wave signal contains a frequency component obtained by multiplying the number of blades of the first rotor and the rotational frequency…"

(source: US 10,054,481 B2 specification, background.)

That is the claimed first/second abnormality taxonomy, pre-dating the '599 priority date, and derived from JP H08-166330 A (a reference cited on the face of both patents). Add JP 2005-062154 A, which computes a frequency from rotational speed and compares it to the measured frequency to identify which part is abnormal. A POSITA combining these with Ground 3 arrives directly at claim 6's subject matter.

Motivation: order/rotational-frequency analysis (and blade-count/order harmonics) is the standard method for localizing machinery faults; the rotor tooth-count ratio relationship (T1×b = T2×a) is an arithmetic consequence of intermeshing rotors, not an inventive insight. Reasonable expectation of success: the cited art already reports doing it.

7. The motivation-to-combine case in one place

A POSITA (mechanical/electrical engineer with ~2–3 years in industrial condition monitoring and AE) would combine the above because:

8. Where the patent might argue back — and why I assess those arguments as weak

1. "Non-obvious because AE had never been applied to this detection task (compressor rotor contact)." Rejected: JP H08-166330 A and the art quoted in US 10,054,481's background apply AE elastic-wave analysis squarely to male/female rotor contact and rotor/casing contact — identically the claimed application.
2. "Secondary considerations / unexpected results." None is apparent on the record. The stated benefit ("reduce redundant operational steps") is the ordinary consequence of event-triggered analysis, not an unexpected result. No evidence of commercial success attributable to the claim, licensing, or copying is in the material reviewed.
3. Criticality of the 100 kHz–1 MHz band. The band appears only in the specification, not in the claims — and in any event AE sensors natively operate there (the art discloses bands from ~20 kHz–900 kHz; US 2016/0055737 gives 5 kHz–900 kHz).
4. "Teaching away." None found. PTL2's use of the audible band is not a disparagement of the ultrasonic band; the '599 spec argues only that PTL2 is insufficient, which is a motivation to improve, not a teaching away.
5. No express "combine" statement. Under KSR, an express combination teaching is unnecessary where the elements are known and the combination performs no more than their predictable functions.

9. Residual uncertainties (candor)

1. I did not open the PTAB/EPO prosecution record for the '599 itself. The "Y" citation I rely on is from the EP family member EP 3 193 155 A1; I retrieved only a snippet of its search-report table, not the full report, and I did not obtain the U.S. examiner's own reasoned rejections or the applicant's remarks. A file-wrapper review (USPTO Patent Center for application 15/328,308) is the single highest-value next step — it would reveal which grounds the examiner actually made and how they were overcome.
2. JP H08-043193 A, JP H08-166330 A, and JP 2010-234403 A were not retrieved in full. My characterizations of the H08 references are inferred from secondary sources (the face citations and US 10,054,481's background) and should be verified before being asserted formally.
3. US 2016/0055737 A1's prior-art date rests on § 102(a)(2) via its 2013-04-08 provisional; it is only prior art for subject matter the provisional actually supports (compare Dynamic Drinkware v. National Graphics). Its publication date (2016-02-25) is after the '599 priority date.
4. US 2019/0049413 A1 is cited only as state-of-the-art evidence, not as prior art; I did not verify its effective filing date.
5. US 2014/0288880 and US 2014/0298099 are dated very near or after the 2014-09-12 critical date; I therefore relied on their earlier-filed counterparts (WO 2012/088707 A1 for the latter) where possible.
6. Claim scope. The claim set reproduced in the record is prose, not a clean numbered list (a limitation flagged in the earlier "Patent summary"). If a claim I did not see recites, e.g., a specific quantitative relationship for Th1 or a specific sensor-mounting geometry, the analysis should be re-run against that text.

10. Bottom line

On the art of record and the surrounding state of the art, claims 1–6 and the machine/method independents are, in my assessment, obvious under § 103 as combinations of (i) an AE/ultrasonic rotating-machine or bearing-diagnosis reference — JP 2002-181038 A, EP 0209862, US 4,790,190, or US 5,072,611; (ii) the frequency-domain/threshold teachings of JP 2001-074616 A (PTL2) and JP 3325448 B2; (iii) the audible-band, normal-vs-current spectral comparison of JP 2002-098585 A, US 2016/0055737 A1, or WO 2012/088707 A1; and (iv) for the rotor-contact taxonomy, JP H08-166330 A and JP 2005-062154 A. Each combination is supported by an articulated motivation (the applicant's own admission of PTL2's masked-band problem; the known incipient-fault sensitivity of ultrasound; the complementary sensitivities of AE and accelerometer; the standard normal-signature-baseline paradigm), and each is supported by a reasonable expectation of success given the maturity of AE condition monitoring well before 2014. The EPO's "Y" citation of JP 2002-181038 A against the corresponding European claim set is independent, objective corroboration of that conclusion.