Capstone: a maintenance audit and an acceptance test day

The term ended with one week that brought together everything Tamsin had learned on Route 7. On Monday and Tuesday, Hugh Mbeki audited the route's maintenance control programs and records, as he did on every Northcrest route once a year and as a Technical Safety BC safety officer might at any time. On Thursday, the new car at Larchmont East, installed over the spring by Northcrest's installation crew, had its acceptance inspection with Camille Draper, the Technical Safety BC safety officer for the area. Lena put Tamsin in charge of preparing for both, with Ray beside her and Priya on the Larchmont paperwork. “The audit asks whether we have been doing the work,” Lena said. “The acceptance asks whether a new car is fit to start. They are the same question at two ends of an elevator's life.” This lesson follows that week from start to finish.

Monday: the audit plan

Hugh's audit followed the questions he had given Tamsin in lesson 2, applied to every unit on the route, with a closer sample of five units for which he would check the equipment as well as the records. For each unit he scored four areas: the program (specific, justified, reviewed), the records (complete, specific, on site and in the system, in agreement), the tests (done on time, with values, comparable) and the equipment (consistent with the records). The sample was Tillicum car 1, Seaview car 2, Birchwood car 2, Holloway car 1 and Dunmore Square, chosen because they had the most callbacks, the most complicated equipment or the oldest records.

Tamsin's preparation was the weekly records hour of lesson 11, done for every unit at once. She reconciled the site logs with the office system, closed out or explained every open item, checked every unit's test dates against the tracker from lesson 6 and checked every program's references against the maker's current procedures. She did not try to make the records look better than they were. Where something was missing or late, the record said so and said what had been done about it.

Monday and Tuesday: the findings

Hugh worked through the binders and the system on Monday and the five sample units on Tuesday, with Tamsin and Ray showing him the equipment. His findings were a mixture of things the term had already found and fixed, and a few that were new.

The Birchwood finding was the one that mattered most. With the oil additions totalled, car 2 had needed about 6 L (1.6 gal) more oil over the year than the year before, with no matching oil in its drip pan. Car 2's cylinder was the same buried, unprotected type as car 1's, whose three-year leakage test had passed in lesson 5, but car 2's own test was not due until the next year. Hugh's recommendation was to do car 2's leakage test now rather than wait, and Lena agreed. Tamsin ran it on the Wednesday with the same method as in lesson 5, with temperature readings, and the unexplained loss was inside the tolerance; the extra oil turned out to have been lost through a weeping fitting on the valve body, hidden behind the power unit's cover and wiped clean on every visit by someone who had never recorded it. The fitting was replaced and the oil record annotated. The result was a small repair, but the process was the point: an audit had found a trend that the visits had missed.

# Audit scores for Route 7 (0 to 5 in each area; sample units also scored on the equipment)
scores = {   # unit: (program, records, tests, equipment or None if not sampled)
    "Tillicum 1":  (3, 4, 5, 4), "Tillicum 2":  (5, 5, 5, None), "Marlowe 1":   (4, 5, 5, None),
    "Larchmont 1": (4, 4, 5, None), "Ashgrove 1": (5, 5, 5, None), "Dunmore 1":   (5, 3, 5, 5),
    "Holloway 1":  (5, 3, 5, 5), "Holloway 2":   (5, 5, 5, None), "Birchwood 1": (5, 5, 5, None),
    "Birchwood 2": (5, 3, 5, 4), "Pemberton 1":  (5, 4, 3, None), "Seaview 1":   (5, 5, 5, None),
    "Seaview 2":   (5, 5, 5, 4), "Bakery":       (5, 5, 5, None),
}
areas = ["program", "records", "tests", "equipment"]
for i, area in enumerate(areas):
    vals = [s[i] for s in scores.values() if s[i] is not None]
    low = [u for u, s in scores.items() if s[i] is not None and s[i] <= 3]
    print(f"{area:9s}: average {sum(vals) / len(vals):.1f} of 5; units at 3 or below: {', '.join(low) or 'none'}")
overall = [x for s in scores.values() for x in s if x is not None]
print(f"route overall: {100 * sum(overall) / (5 * len(overall)):.0f} %")
weak = max(areas, key=lambda a: sum(1 for s in scores.values() if s[areas.index(a)] is not None and s[areas.index(a)] <= 3))
print("area to improve first:", weak)

program  : average 4.7 of 5; units at 3 or below: Tillicum 1
records  : average 4.4 of 5; units at 3 or below: Dunmore 1, Holloway 1, Birchwood 2
tests    : average 4.9 of 5; units at 3 or below: Pemberton 1
equipment: average 4.4 of 5; units at 3 or below: none
route overall: 92 %
area to improve first: records

The route scored well overall, and the scores pointed to records as the area to improve, which matched everything Tamsin had seen in the Friday hours. Hugh's report to Lena listed the eight findings with their actions, owners and dates, and made one route-wide recommendation: a photograph of every posted test sheet and every oil record page at every visit, attached to the visit in the office system, so that a missing entry would show up in the office within a day instead of at the next audit.

Wednesday: getting Larchmont East ready

Larchmont East was a six-storey rental building beside the original Larchmont, with one machine-room-less car on coated steel belts from the same maker, rated 1,135 kg (2,500 lb) at 1.0 m/s (200 ft/min). Northcrest's installation crew had finished the work two weeks earlier and run the full set of pre-tests, and Priya's readiness sheet, built like the one in lesson 10 but four pages long, had been closed item by item. On Wednesday afternoon, after the Birchwood test, Tamsin and Ray walked the whole installation once more as the inspector would: the control cabinet and its disconnects beside the top landing, the hoistway on inspection from top to bottom, the pit with its refuge space and signs, the car and the six landings. They found one thing: a pit sign stating the refuge space dimensions had been fixed to the wrong wall, where the counterweight screen hid it. It was moved before they left.

The documents were in a binder on top of the control cabinet: the permit, the product certificates for the controller, the machine and brake, the governor, the belts and their monitoring device, the door operator and the light curtain, the maker's installation and adjustment instructions, the as-built wiring diagrams, the pre-test records with every value, the data plates' text, and the new car's maintenance control program, written from Larchmont car 1's program with its reviewed door intervals and with extra first-year tasks for door adjustment. Because Larchmont East was permitted under British Columbia's adopted edition, its emergency communication was a two-way voice system with a monitored line, tested and recorded.

Thursday: the acceptance

Camille Draper arrived at eight, read the binder for half an hour, and asked to see the control cabinet, the car top and the pit before any test. Then the tests began, run by Ray and Tamsin with Jonah moving the weights, in the order the pre-tests had used: the governor and its overspeed switch, using the maker's remote trip and reset; the safeties with rated load at rated speed; the oil buffers struck with rated load at the speed the procedure required; the brake holding 125 % of rated load and then each half of the brake on its own; ascending car overspeed and unintended movement in both directions; the belt monitoring device's test function; the terminal and emergency terminal devices; the doors, door protection and kinetic energy; firefighters' emergency operation with the building's fire alarm technician; the emergency communication and car lighting; and the battery rescue operation that brings the car to a landing in a power failure.

# Larchmont East car E1 acceptance results against the program's criteria
results = [  # (test, measured, low limit or None, high limit or None, unit)
    ("governor trip speed",            1.33,  1.15, 1.42, "m/s"),
    ("safety stopping distance",       0.21,  0.1, 0.4, "m"),
    ("platform tilt after safety stop", 1.1, None, 3.0, "%"),
    ("car buffer return time",          38,  None, 90, "s"),
    ("brake hold, 125 % load, creep",   0,   None, 0, "m"),
    ("brake half A stop from speed",   0.71, None, 1.2, "m"),
    ("brake half B stop from speed",   0.74, None, 1.2, "m"),
    ("unintended movement, up empty",  0.29, None, 1.22, "m"),
    ("unintended movement, down loaded", 0.24, None, 1.22, "m"),
    ("door average kinetic energy",    6.2, None, 10, "J"),
    ("rescue operation to landing",     95,  None, 300, "s"),
]
fails = 0
for name, x, lo, hi, u in results:
    ok = (lo is None or x >= lo) and (hi is None or x <= hi)
    fails += not ok
    rng = f"{lo if lo is not None else ''}..{hi if hi is not None else ''}"
    print(f"{name:33s} {x:>7} {u:6s} limits {rng:12s} {'pass' if ok else 'FAIL'}")
print(f"{len(results) - fails} of {len(results)} tests passed")

governor trip speed                  1.33 m/s    limits 1.15..1.42   pass
safety stopping distance             0.21 m      limits 0.1..0.4     pass
platform tilt after safety stop       1.1 %      limits ..3.0        pass
car buffer return time                 38 s      limits ..90         pass
brake hold, 125 % load, creep           0 m      limits ..0          pass
brake half A stop from speed         0.71 m      limits ..1.2        pass
brake half B stop from speed         0.74 m      limits ..1.2        pass
unintended movement, up empty        0.29 m      limits ..1.22       pass
unintended movement, down loaded     0.24 m      limits ..1.22       pass
door average kinetic energy           6.2 J      limits ..10         pass
rescue operation to landing            95 s      limits ..300        pass
11 of 11 tests passed

Every test passed, and the values matched the pre-test records closely, which Camille noted with approval: a pre-test sheet that matches the acceptance is good evidence that the contractor tests honestly. She raised two findings. The first was the pit sign, which she had found on the right wall and said nothing about until Tamsin mentioned it had been moved the day before; she smiled and wrote it down as corrected. The second was real: the governor's tripping speed was marked on its data tag in the maker's factory units, while the code requires the tag to show the tripping speed and rope details, and the tag's figure did not match the setting on the test record because it had been rounded differently. It was a minor item, to be corrected within thirty days with a new tag and a photograph.

At four o'clock Camille signed off the acceptance with the one open minor item. Technical Safety BC issued the operating permit the following morning. The car went into service that afternoon under its new maintenance control program, with its first visit scheduled for two weeks later and its acceptance values entered as the baseline for every test to come.

Friday: closing the week

On Friday Tamsin spent the records hour on the week itself. Hugh's audit findings went into the tracker with their owners and dates. Larchmont East's acceptance record, with every value, instrument and person, went into the binder and the system; the governor tag was ordered; the route model from lesson 1 was updated with the new car's actual first-month hours. Lena read the audit report and the acceptance record together and wrote a short note at the bottom of both: the route had found and fixed most of its own problems before the audit, and a new car had passed its acceptance at the first attempt. Both, she wrote, came from the same habit of measuring, recording and acting on what the records said.

The next term would bring Tillicum car 1's modernization, which ELV-430 follows from the first day of demolition to its own acceptance test day: the old car's last five-year test from lesson 7 would become the starting point for the new equipment's baseline.