关于「东西方原典考辨:沈括论漏刻与贝塞尔的个人方程」,《出生时刻误差与交界翻转》给出的定量结论是什么?

结论:第七章考辨沈括论漏刻与贝塞尔的个人方程,并说明在中国西部真太阳时这一个约定能改变 0.9932 的出生时辰。

数据口径:本书只分析排盘规则的结构、误差传播与概率,不做任何个人占断,不宣称任何体系有预测效力;书中参数为示意,全部计算由书中程序生成。 第一章说明,示意模型中凭记忆报告的时刻有 0.2782 恰好落在时辰分界上,不取整时只有 0.0083。

English version

对照数据表

下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库与结论吻合的打印值: 0.9932

equation of time on the 15th of each month (minutes)
month1 2 3 4 5 6
E-9.3 -14.6 -9.1 0.3 3.7 -0.6
month7 8 9 10 11 12
E-5.8 -3.8 5.9 15.2 14.7 3.3
max +16.5 min on day 304, min -14.6 min on day 44
longest apparent day: +28.2 s (day 356); shortest-22.3 s (day 258)
day length minus 24 h (s): total = eccentric + oblique part
day 172: +12.6-7.8 +20.4
day 356: +28.2 = +7.8 +20.4
city lon offset(min) share of double hours changed
Shanghai121.5 6.0 0.0702
Beijing116.4 -14.4 0.1222
Xi'an108.9 -44.4 0.3700
Chengdu104.1 -63.6 0.5300
Lhasa91.1 -115.6 0.9492
Urumqi87.6 -129.6 0.9932
查看完整验算程序
# Chapter 7. p71: Shen Kuo's clepsydra puzzle as the equation of time.
# Shen Kuo noted that around the winter solstice a solar day ran longer
# than 100 ke and around the summer solstice shorter. We use the common
# approximation of the equation of time (apparent minus mean solar
# time, minutes): E = 9.87 sin 2B - 7.53 cos B - 1.5 sin B with
# B = 360 (N - 81) / 365 degrees, N the day of the year (accuracy about
# a minute, enough for scale). The length of the apparent solar day
# exceeds 24 h by -dE/dN (in seconds, times 60); we split it into the
# part from the eccentric orbit (cos B and sin B terms) and the part
# from the tilt of the ecliptic (sin 2B term). Then the clock offset
# for places on UTC+8 standard time (zone meridian 120 E), and the
# share of uniformly spread births whose double hour differs between
# clock time and local apparent time, min(|offset + E| / 120, 1),
# averaged over the year.
import math


def eot(n):
    b = math.radians(360.0 * (n - 81) / 365.0)
    return 9.87 * math.sin(2 * b) - 7.53 * math.cos(b) - 1.5 * math.sin(b)


DAYS = [(n, eot(n)) for n in range(1, 366)]
print("equation of time on the 15th of each month (minutes)")
months = [15 + 30 * k + (k + 1) // 2 for k in range(12)]
for half in (0, 6):
    print("  month " + " ".join("%6d" % (k + 1)
                                for k in range(half, half + 6)))
    print("  E     " + " ".join("%6.1f" % eot(months[k])
                                for k in range(half, half + 6)))
hi = max(DAYS, key=lambda d: d[1])
lo = min(DAYS, key=lambda d: d[1])
print("max %+.1f min on day %d, min %+.1f min on day %d"
      % (hi[1], hi[0], lo[1], lo[0]))
excess = [(n, -(eot(n + 0.5) - eot(n - 0.5)) * 60.0)
          for n in range(1, 366)]
lng = max(excess, key=lambda d: d[1])
sht = min(excess, key=lambda d: d[1])
print("longest apparent day: +%.1f s (day %d); shortest: %.1f s (day %d)"
      % (lng[1], lng[0], sht[1], sht[0]))


def part_ecc(n):
    b = math.radians(360.0 * (n - 81) / 365.0)
    return -7.53 * math.cos(b) - 1.5 * math.sin(b)


def part_obl(n):
    b = math.radians(360.0 * (n - 81) / 365.0)
    return 9.87 * math.sin(2 * b)


print("day length minus 24 h (s): total = eccentric + oblique part")
for n in (172, 356):
    tot = dict(excess)[n]
    ecc = -(part_ecc(n + 0.5) - part_ecc(n - 0.5)) * 60.0
    obl = -(part_obl(n + 0.5) - part_obl(n - 0.5)) * 60.0
    print("  day %d: %+.1f = %+.1f %+.1f" % (n, tot, ecc, obl))
CITIES = [("Shanghai", 121.5), ("Beijing", 116.4), ("Xi'an", 108.9),
          ("Chengdu", 104.1), ("Lhasa", 91.1), ("Urumqi", 87.6)]
print("city      lon   offset(min)  share of double hours changed")
for name, lon in CITIES:
    off = 4.0 * (lon - 120.0)
    share = math.fsum(min(abs(off + e) / 120.0, 1.0) for _, e in DAYS)
    print("%-9s %5.1f %10.1f %18.4f" % (name, lon, off, share / 365.0))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《出生时刻记忆误差传播与鲁棒定盘、交界时辰盘面翻转概率与最少前事反推题数母典》第 7 章「东西方原典考辨:沈括论漏刻与贝塞尔的个人方程」(版本 1.0.0). 5D Library. DOI 待发布

DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-284 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026284,
  author    = {Fan, Yafu},
  title     = {Error Propagation from Remembered Birth Times and Robust Chart Fixing: Flip Probabilities at Double-Hour Boundaries and the Fewest Past-Event Questions},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论