IndPenSim-Golden-Operator — Golden batch and multiway PCA

+------------------------------------------------------------------------------
| Program Name:         step_batch.py
| Program Version:      2026-09-10 18:41:38
| Date:                 2026-09-11
| Time:                 08:50:57
| Input 1 - Raw CSV:    indpensim_process.csv  (100 batches, 23 tags)
| Input 2 - Glossary:   indpensim_schema.csv
| Input 3 - Time stamp: aligned onto 100 points of batch progress
+------------------------------------------------------------------------------

In short. 100 batches were compared against a golden batch built from 20 of them. 66 sit more than 10 sigma away at their worst point, and 10 of the 10 batches with a recorded fault were caught.

1. The input data

Where it came from

Three steps, so any number in this report can be traced back to the file it was downloaded from. The links open the file or its folder on this machine.

What	File
1. Raw data, as downloaded	100_Batches_IndPenSim_V3.csv 2.6 GB the original, kept on the drive and never edited /Volumes/YN-4T/WIP/CTRL-Designer/B-CDL-Data-Analysis/A-Raw-Data/IndPenSim	open the folder
2. Program that converted it	make_indpensim.py 8 kB run this again to rebuild the analysis file /Users/yahyanazer/Dropbox/__C_2026_Work/CTRL-Designer/B-CDL-Data-Analysis/B-Engines	open the folder
3. Data used for this analysis	indpensim_process.csv 19 MB what the numbers in this report were computed from /Users/yahyanazer/Dropbox/__C_2026_Work/CTRL-Designer/B-CDL-Data-Analysis/A-Raw-Data/IndPenSim	open the folder
Batch list	indpensim_batches.csv 5 kB /Users/yahyanazer/Dropbox/__C_2026_Work/CTRL-Designer/B-CDL-Data-Analysis/A-Raw-Data/IndPenSim	open the folder
Glossary	indpensim_schema.csv 1 kB /Users/yahyanazer/Dropbox/__C_2026_Work/CTRL-Designer/B-CDL-Data-Analysis/A-Raw-Data/IndPenSim	open the folder
Data sheet	data-source.md 8 kB what the dataset is, its licence, and its known quirks /Users/yahyanazer/Dropbox/__C_2026_Work/CTRL-Designer/B-CDL-Data-Analysis/A-Raw-Data/IndPenSim	open the folder

What it holds

Batches	100
Tags used	23
Batch length	167 to 290 h (median 230 h)
Alignment	every batch stretched onto 100 points of 0–100% progress, because they run to different lengths
Corridor	mean ± 3 sigma of the reference batches

How the batches were run

Strategy	Batches	Mean length	Mean yield
APC with Raman	30	225 h	3,484,283 kg
faulty	10	230 h	2,600,203 kg
operator controlled	30	229 h	2,833,113 kg
recipe driven	30	229 h	2,912,750 kg

Left out of the analysis: agitator_rpm (constant in every batch), paa_conc_offline (offline, 98% blank), nh3_conc_offline (offline, 98% blank), penicillin_offline (offline, 98% blank), biomass_offline (offline, 98% blank), nh3_shots (constant in every batch), viscosity_offline (offline, 98% blank).

2. The golden batch

Built from 20 reference batches. At every one of the 100 points of progress it holds a mean and a spread for each of the 23 tags. Those 20 batches sit on average 1.1% outside their own corridor, which is the noise floor of the method — nothing can score better than that.

The corridor for every tag is drawn in section 6. Here is what the whole set looks like measured against it:

the same as a picture (for printing)

One point per batch. T² across is movement inside the pattern the good batches share; SPE up is what the model cannot explain at all. Both on log scales.

3. Every batch against the golden batch

Each batch is placed by its worst single point — how far from the golden batch it ever got, in sigma. That is the plain reading of "stays within ± so many sigma". The sustained figure beside it is the 99.9th percentile, which ignores one bad sample and says where the batch really lived.

	Band	Batches	How far	Which
1	Near the golden batch — within 2 sigma	0	none
2	Similar — 2 to 5 sigma a normal batch: it wandered, but never far	27 0 with a recorded fault	worst point 2–5 sigma sustained 4 sigma (median)	Batch 4 - recipe driven, Batch 8 - recipe driven, Batch 10 - recipe driven, Batch 16 - recipe driven, Batch 22 - recipe driven, Batch 24 - recipe driven, Batch 31 - operator controlled, Batch 32 - operator controlled, Batch 33 - operator controlled, Batch 34 - operator controlled, Batch 35 - operator controlled, Batch 36 - operator controlled, Batch 37 - operator controlled, Batch 38 - operator controlled, Batch 39 - operator controlled, Batch 40 - operator controlled, Batch 41 - operator controlled, Batch 42 - operator controlled, Batch 43 - operator controlled, Batch 44 - operator controlled, Batch 45 - operator controlled, Batch 46 - operator controlled …
3	Outside — 5 to 10 sigma clearly different from the golden batch at some point	7 0 with a recorded fault	worst point 5–10 sigma sustained 5 sigma (median)	Batch 2 - recipe driven, Batch 12 - recipe driven, Batch 14 - recipe driven, Batch 20 - recipe driven, Batch 25 - recipe driven, Batch 26 - recipe driven, Batch 54 - operator controlled
4	Far from it — over 10 sigma a different batch altogether - something happened	66 10 with a recorded fault	worst point 10–183 sigma sustained 36 sigma (median)	Batch 1 - recipe driven, Batch 3 - recipe driven, Batch 5 - recipe driven, Batch 6 - recipe driven, Batch 7 - recipe driven, Batch 9 - recipe driven, Batch 11 - recipe driven, Batch 13 - recipe driven, Batch 15 - recipe driven, Batch 17 - recipe driven, Batch 18 - recipe driven, Batch 19 - recipe driven, Batch 21 - recipe driven, Batch 23 - recipe driven, Batch 27 - recipe driven, Batch 28 - recipe driven, Batch 29 - recipe driven, Batch 30 - recipe driven, Batch 51 - operator controlled, Batch 52 - operator controlled, Batch 53 - operator controlled, Batch 55 - operator controlled …

A batch holds 23 tags x 100 points = 2,300 readings. The largest of that many draws from a normal distribution is about 3.5 sigma by chance alone, so the first band is expected to be thinly populated or empty even for perfect batches. Change the edges in the window if you want bands that split your own data more usefully.

4. What caused the deviation

The 66 batches beyond 10 sigma, worst first. "Worst tag" is the single tag furthest from the golden batch, and "at" is where in the batch that happened.

Batch	Fault recorded	Worst point	Worst tag	at	first left	time outside
100	faulty	yes	183 sigma	aeration_rate	46%	9%	5.5%
94	faulty	yes	183 sigma	aeration_rate	46%	9%	7.2%
51	operator controlled	no	178 sigma	acid_flow_rate	86%	1%	1.2%
59	operator controlled	no	172 sigma	acid_flow_rate	85%	7%	2.9%
67	APC with Raman	no	166 sigma	ph	48%	2%	20.0%
97	faulty	yes	162 sigma	acid_flow_rate	77%	9%	4.9%
91	faulty	yes	156 sigma	substrate_conc	46%	5%	19.5%
9	recipe driven	no	152 sigma	acid_flow_rate	61%	10%	4.6%
93	faulty	yes	135 sigma	temperature	42%	1%	3.6%
99	faulty	yes	135 sigma	acid_flow_rate	77%	3%	12.3%
95	faulty	yes	115 sigma	acid_flow_rate	63%	10%	12.8%
3	recipe driven	no	113 sigma	acid_flow_rate	98%	1%	12.0%
92	faulty	yes	111 sigma	ph	69%	7%	1.9%
98	faulty	yes	98 sigma	ph	69%	6%	1.4%
76	APC with Raman	no	78 sigma	heat_water_flow	15%	2%	1.8%
7	recipe driven	no	75 sigma	ph	86%	1%	3.2%
15	recipe driven	no	67 sigma	head_pressure	21%	2%	10.3%
5	recipe driven	no	67 sigma	head_pressure	21%	2%	10.3%
68	APC with Raman	no	66 sigma	paa_flow	38%	2%	2.9%
70	APC with Raman	no	66 sigma	paa_flow	38%	5%	2.7%
29	recipe driven	no	66 sigma	paa_flow	14%	1%	12.0%
17	recipe driven	no	65 sigma	water_injection	26%	2%	5.6%
75	APC with Raman	no	62 sigma	heat_water_flow	14%	2%	10.6%
28	recipe driven	no	59 sigma	acid_flow_rate	78%	61%	0.7%
23	recipe driven	no	56 sigma	paa_flow	15%	3%	4.2%
53	operator controlled	no	52 sigma	water_injection	38%	0%	9.8%
64	APC with Raman	no	47 sigma	paa_flow	39%	0%	2.3%
61	APC with Raman	no	46 sigma	dumped_broth_flow	54%	3%	6.4%
85	APC with Raman	no	46 sigma	dumped_broth_flow	76%	6%	5.6%
55	operator controlled	no	46 sigma	dumped_broth_flow	46%	0%	1.0%
1	recipe driven	no	46 sigma	dumped_broth_flow	76%	0%	2.0%
19	recipe driven	no	46 sigma	dumped_broth_flow	64%	11%	1.6%
13	recipe driven	no	46 sigma	dumped_broth_flow	46%	11%	1.5%
27	recipe driven	no	46 sigma	dumped_broth_flow	48%	11%	2.0%
71	APC with Raman	no	46 sigma	dumped_broth_flow	48%	6%	3.2%
81	APC with Raman	no	46 sigma	dumped_broth_flow	60%	7%	2.1%
73	APC with Raman	no	46 sigma	dumped_broth_flow	42%	5%	9.2%
57	operator controlled	no	46 sigma	dumped_broth_flow	99%	34%	0.5%
87	APC with Raman	no	46 sigma	dumped_broth_flow	46%	0%	2.6%
83	APC with Raman	no	46 sigma	dumped_broth_flow	54%	5%	3.3%
77	APC with Raman	no	46 sigma	dumped_broth_flow	50%	10%	3.3%
69	APC with Raman	no	46 sigma	dumped_broth_flow	59%	10%	4.2%
65	APC with Raman	no	41 sigma	paa_flow	10%	0%	1.6%
62	APC with Raman	no	41 sigma	paa_flow	15%	10%	1.7%
80	APC with Raman	no	39 sigma	paa_flow	15%	10%	1.9%
74	APC with Raman	no	39 sigma	paa_flow	15%	5%	1.4%
78	APC with Raman	no	39 sigma	paa_flow	12%	1%	3.0%
89	APC with Raman	no	38 sigma	paa_flow	39%	5%	3.6%
90	APC with Raman	no	38 sigma	paa_flow	12%	0%	1.6%
84	APC with Raman	no	38 sigma	paa_flow	15%	0%	2.1%
79	APC with Raman	no	37 sigma	paa_flow	15%	10%	2.0%
82	APC with Raman	no	36 sigma	paa_flow	15%	11%	1.8%
88	APC with Raman	no	36 sigma	paa_flow	15%	3%	1.8%
63	APC with Raman	no	34 sigma	paa_flow	15%	11%	1.9%
86	APC with Raman	no	34 sigma	paa_flow	12%	11%	1.9%
72	APC with Raman	no	33 sigma	paa_flow	15%	10%	1.4%
66	APC with Raman	no	32 sigma	paa_flow	12%	3%	3.5%
96	faulty	yes	30 sigma	temperature	36%	33%	2.0%
30	recipe driven	no	26 sigma	acid_flow_rate	58%	1%	1.5%
52	operator controlled	no	22 sigma	heat_water_flow	20%	2%	0.6%
21	recipe driven	no	20 sigma	heat_water_flow	5%	3%	1.8%
11	recipe driven	no	19 sigma	sugar_feed_rate	38%	6%	2.1%
6	recipe driven	no	14 sigma	acid_flow_rate	37%	6%	2.4%
60	operator controlled	no	12 sigma	temperature	24%	18%	0.4%
18	recipe driven	no	11 sigma	acid_flow_rate	40%	17%	1.3%
58	operator controlled	no	10 sigma	acid_flow_rate	35%	5%	0.2%

The milder bands

Similar — 2 to 5 sigma — 27 batches, 0 with a recorded fault. Furthest tag most often: acid_flow_rate (8), temperature (4), heat_water_flow (3). The worst point typically falls at 57% of the batch.
Outside — 5 to 10 sigma — 7 batches, 0 with a recorded fault. Furthest tag most often: temperature (2), base_flow_rate (2), heat_water_flow (2). The worst point typically falls at 13% of the batch.

Batch by batch

Each chart holds every tag of that batch on one axis, measured in sigma from the golden batch. The green line at zero is the golden trajectory, the dashed red lines are the ±3 sigma corridor. The tags driving the batch are bold, the rest are faint behind them so you can still see whether the whole batch moved or only a few tags.

1. Batch 67 - APC with Raman does not look like a golden batchseverity 10
APC with Raman · no fault recorded for this batch · first left the corridor at 2% of the way through

the same as a picture (for printing)

What the chart shows. ph carries most of it, running 166 sigma below the golden batch at 47% of the way through; 20 of 23 tags leave the corridor at some point; the first departure is at 2%, almost from the start; a large share of the batch sits outside, so this batch did not recover.

The numbers. SPE 305387 against a limit of 520 (587.2x); T² 25 against a limit of 18 (1.4x); 20.0% of the batch sits outside the golden corridor.

Driven by: ph (51%), head_pressure (12%), water_injection (11%), oil_flow (9%), aeration_rate (7%)

2. Batch 91 - faulty does not look like a golden batchseverity 10
faulty · a fault IS recorded for this batch · first left the corridor at 5% of the way through

the same as a picture (for printing)

What the chart shows. substrate_conc carries most of it, running 156 sigma above the golden batch at 45% of the way through; 19 of 23 tags leave the corridor at some point; the first departure is at 5%, almost from the start; a large share of the batch sits outside, so this batch did not recover.

The numbers. SPE 162277 against a limit of 520 (312.0x); T² 36 against a limit of 18 (2.0x); 19.5% of the batch sits outside the golden corridor.

Driven by: substrate_conc (59%), acid_flow_rate (33%), dumped_broth_flow (3%), sugar_feed_rate (2%), temperature (0%)

3. Batch 94 - faulty does not look like a golden batchseverity 10
faulty · a fault IS recorded for this batch · first left the corridor at 9% of the way through

the same as a picture (for printing)

What the chart shows. aeration_rate carries most of it, running 183 sigma below the golden batch at 46% of the way through; 14 of 23 tags leave the corridor at some point; the first departure is at 9%, almost from the start.

The numbers. SPE 108798 against a limit of 520 (209.2x); T² 34 against a limit of 18 (1.9x); 7.2% of the batch sits outside the golden corridor.

Driven by: aeration_rate (63%), acid_flow_rate (17%), co2_offgas (6%), cool_water_flow (5%), temperature (3%)

4. Batch 100 - faulty does not look like a golden batchseverity 10
faulty · a fault IS recorded for this batch · first left the corridor at 9% of the way through

the same as a picture (for printing)

What the chart shows. aeration_rate carries most of it, running 183 sigma below the golden batch at 46% of the way through; 16 of 23 tags leave the corridor at some point; the first departure is at 9%, almost from the start.

The numbers. SPE 104893 against a limit of 520 (201.7x); T² 19 against a limit of 18 (1.1x); 5.5% of the batch sits outside the golden corridor.

Driven by: aeration_rate (65%), acid_flow_rate (18%), co2_offgas (9%), o2_offgas (4%), substrate_conc (2%)

5. Batch 99 - faulty does not look like a golden batchseverity 10
faulty · a fault IS recorded for this batch · first left the corridor at 3% of the way through

the same as a picture (for printing)

What the chart shows. acid_flow_rate carries most of it, running 135 sigma above the golden batch at 77% of the way through; 19 of 23 tags leave the corridor at some point; the first departure is at 3%, almost from the start.

The numbers. SPE 89135 against a limit of 520 (171.4x); T² 21 against a limit of 18 (1.2x); 12.3% of the batch sits outside the golden corridor.

Driven by: acid_flow_rate (56%), substrate_conc (20%), co2_offgas (6%), o2_offgas (4%), heat_water_flow (3%)

6. Batch 51 - operator controlled does not look like a golden batchseverity 10
operator controlled · no fault recorded for this batch · first left the corridor at 1% of the way through

the same as a picture (for printing)

What the chart shows. acid_flow_rate carries most of it, running 178 sigma above the golden batch at 86% of the way through; 8 of 23 tags leave the corridor at some point; the first departure is at 1%, almost from the start; only a small part of the batch is outside, so this is a short, sharp deviation rather than a batch that ran wrong throughout.

The numbers. SPE 85332 against a limit of 520 (164.1x); 1.2% of the batch sits outside the golden corridor.

Driven by: acid_flow_rate (90%), dumped_broth_flow (8%), ph (1%), sugar_feed_rate (1%), cool_water_flow (0%)

7. Batch 5 - recipe driven does not look like a golden batchseverity 10
recipe driven · no fault recorded for this batch · first left the corridor at 2% of the way through

the same as a picture (for printing)

What the chart shows. head_pressure carries most of it, running 67 sigma below the golden batch at 21% of the way through; 20 of 23 tags leave the corridor at some point; the first departure is at 2%, almost from the start.

The numbers. SPE 82483 against a limit of 520 (158.6x); 10.3% of the batch sits outside the golden corridor.

Driven by: head_pressure (27%), water_injection (24%), oil_flow (16%), aeration_rate (14%), paa_flow (11%)

8. Batch 15 - recipe driven does not look like a golden batchseverity 10
recipe driven · no fault recorded for this batch · first left the corridor at 2% of the way through

the same as a picture (for printing)

What the chart shows. water_injection carries most of it, running 65 sigma below the golden batch at 26% of the way through; 22 of 23 tags leave the corridor at some point; the first departure is at 2%, almost from the start.

The numbers. SPE 69122 against a limit of 520 (132.9x); 10.3% of the batch sits outside the golden corridor.

Driven by: water_injection (26%), head_pressure (24%), aeration_rate (15%), oil_flow (14%), paa_flow (6%)

9. Batch 29 - recipe driven does not look like a golden batchseverity 10
recipe driven · no fault recorded for this batch · first left the corridor at 1% of the way through

the same as a picture (for printing)

What the chart shows. aeration_rate carries most of it, running 50 sigma above the golden batch at 31% of the way through; 21 of 23 tags leave the corridor at some point; the first departure is at 1%, almost from the start.

The numbers. SPE 66864 against a limit of 520 (128.6x); 12.0% of the batch sits outside the golden corridor.

Driven by: aeration_rate (27%), water_injection (26%), head_pressure (15%), paa_flow (14%), dumped_broth_flow (6%)

10. Batch 3 - recipe driven does not look like a golden batchseverity 10
recipe driven · no fault recorded for this batch · first left the corridor at 1% of the way through

the same as a picture (for printing)

What the chart shows. acid_flow_rate carries most of it, running 113 sigma above the golden batch at 98% of the way through; 23 of 23 tags leave the corridor at some point; the first departure is at 1%, almost from the start.

The numbers. SPE 58789 against a limit of 520 (113.0x); 12.0% of the batch sits outside the golden corridor.

Driven by: acid_flow_rate (25%), water_injection (19%), dumped_broth_flow (16%), aeration_rate (12%), head_pressure (12%)

11. Batch 95 - faulty does not look like a golden batchseverity 10
faulty · a fault IS recorded for this batch · first left the corridor at 10% of the way through

the same as a picture (for printing)

What the chart shows. acid_flow_rate carries most of it, running 115 sigma above the golden batch at 63% of the way through; 20 of 23 tags leave the corridor at some point; the first departure is at 10%, early in the batch.

The numbers. SPE 49801 against a limit of 520 (95.8x); T² 21 against a limit of 18 (1.2x); 12.8% of the batch sits outside the golden corridor.

Driven by: acid_flow_rate (62%), sugar_feed_rate (24%), dumped_broth_flow (5%), ph (1%), oxygen_uptake (1%)

12. Batch 9 - recipe driven does not look like a golden batchseverity 10
recipe driven · no fault recorded for this batch · first left the corridor at 10% of the way through

the same as a picture (for printing)

What the chart shows. acid_flow_rate carries most of it, running 152 sigma above the golden batch at 61% of the way through; 14 of 23 tags leave the corridor at some point; the first departure is at 10%, early in the batch.

The numbers. SPE 47267 against a limit of 520 (90.9x); 4.6% of the batch sits outside the golden corridor.

Driven by: acid_flow_rate (82%), dumped_broth_flow (14%), ph (1%), paa_flow (1%), heat_water_flow (1%)

13. Batch 92 - faulty does not look like a golden batchseverity 10
faulty · a fault IS recorded for this batch · first left the corridor at 7% of the way through

the same as a picture (for printing)

What the chart shows. ph carries most of it, running 111 sigma below the golden batch at 69% of the way through; 7 of 23 tags leave the corridor at some point; the first departure is at 7%, almost from the start; only a small part of the batch is outside, so this is a short, sharp deviation rather than a batch that ran wrong throughout.

The numbers. SPE 46354 against a limit of 520 (89.1x); 1.9% of the batch sits outside the golden corridor.

Driven by: ph (97%), base_flow_rate (1%), paa_flow (1%), temperature (0%), heat_water_flow (0%)

14. Batch 97 - faulty does not look like a golden batchseverity 10
faulty · a fault IS recorded for this batch · first left the corridor at 9% of the way through

the same as a picture (for printing)

What the chart shows. acid_flow_rate carries most of it, running 162 sigma above the golden batch at 77% of the way through; 16 of 23 tags leave the corridor at some point; the first departure is at 9%, almost from the start.

The numbers. SPE 45875 against a limit of 520 (88.2x); 4.9% of the batch sits outside the golden corridor.

Driven by: acid_flow_rate (66%), dumped_broth_flow (15%), sugar_feed_rate (13%), ph (1%), cool_water_flow (1%)

15. Batch 59 - operator controlled does not look like a golden batchseverity 10
operator controlled · no fault recorded for this batch · first left the corridor at 7% of the way through

the same as a picture (for printing)

What the chart shows. acid_flow_rate carries most of it, running 172 sigma above the golden batch at 85% of the way through; 16 of 23 tags leave the corridor at some point; the first departure is at 7%, almost from the start; only a small part of the batch is outside, so this is a short, sharp deviation rather than a batch that ran wrong throughout.

The numbers. SPE 42270 against a limit of 520 (81.3x); 2.9% of the batch sits outside the golden corridor.

Driven by: acid_flow_rate (81%), dumped_broth_flow (10%), heat_water_flow (4%), ph (1%), sugar_feed_rate (1%)

16. Batch 17 - recipe driven does not look like a golden batchseverity 10
recipe driven · no fault recorded for this batch · first left the corridor at 2% of the way through

the same as a picture (for printing)

What the chart shows. water_injection carries most of it, running 65 sigma below the golden batch at 26% of the way through; 20 of 23 tags leave the corridor at some point; the first departure is at 2%, almost from the start.

The numbers. SPE 39735 against a limit of 520 (76.4x); 5.6% of the batch sits outside the golden corridor.

Driven by: water_injection (25%), head_pressure (23%), oil_flow (15%), aeration_rate (13%), paa_flow (7%)

17. Batch 7 - recipe driven does not look like a golden batchseverity 10
recipe driven · no fault recorded for this batch · first left the corridor at 1% of the way through

the same as a picture (for printing)

What the chart shows. ph carries most of it, running 75 sigma below the golden batch at 86% of the way through; 6 of 23 tags leave the corridor at some point; the first departure is at 1%, almost from the start.

The numbers. SPE 36351 against a limit of 520 (69.9x); 3.2% of the batch sits outside the golden corridor.

Driven by: ph (95%), base_flow_rate (2%), paa_flow (1%), temperature (0%), vessel_volume (0%)

18. Batch 75 - APC with Raman does not look like a golden batchseverity 10
APC with Raman · no fault recorded for this batch · first left the corridor at 2% of the way through

the same as a picture (for printing)

What the chart shows. paa_flow carries most of it, running 32 sigma above the golden batch at 13% of the way through; 21 of 23 tags leave the corridor at some point; the first departure is at 2%, almost from the start.

The numbers. SPE 31381 against a limit of 520 (60.3x); 10.6% of the batch sits outside the golden corridor.

Driven by: paa_flow (30%), heat_water_flow (13%), head_pressure (11%), ph (10%), water_injection (9%)

19. Batch 98 - faulty does not look like a golden batchseverity 10
faulty · a fault IS recorded for this batch · first left the corridor at 6% of the way through

the same as a picture (for printing)

What the chart shows. ph carries most of it, running 98 sigma below the golden batch at 69% of the way through; 6 of 23 tags leave the corridor at some point; the first departure is at 6%, almost from the start; only a small part of the batch is outside, so this is a short, sharp deviation rather than a batch that ran wrong throughout.

The numbers. SPE 25741 against a limit of 520 (49.5x); 1.4% of the batch sits outside the golden corridor.

Driven by: ph (96%), paa_flow (1%), base_flow_rate (1%), generated_heat (1%), temperature (0%)

20. Batch 93 - faulty does not look like a golden batchseverity 10
faulty · a fault IS recorded for this batch · first left the corridor at 1% of the way through

the same as a picture (for printing)

What the chart shows. temperature carries most of it, running 135 sigma above the golden batch at 42% of the way through; 17 of 23 tags leave the corridor at some point; the first departure is at 1%, almost from the start.

The numbers. SPE 23171 against a limit of 520 (44.6x); 3.6% of the batch sits outside the golden corridor.

Driven by: temperature (86%), generated_heat (3%), heat_water_flow (2%), base_flow_rate (1%), cool_water_flow (1%)

21. Batch 71 - APC with Raman does not look like a golden batchseverity 10
APC with Raman · no fault recorded for this batch · first left the corridor at 6% of the way through

the same as a picture (for printing)

What the chart shows. paa_flow carries most of it, running 31 sigma above the golden batch at 12% of the way through; 16 of 23 tags leave the corridor at some point; the first departure is at 6%, almost from the start.

The numbers. SPE 20897 against a limit of 520 (40.2x); 3.2% of the batch sits outside the golden corridor.

Driven by: paa_flow (43%), dumped_broth_flow (41%), heat_water_flow (7%), acid_flow_rate (3%), temperature (1%)

22. Batch 73 - APC with Raman does not look like a golden batchseverity 10
APC with Raman · no fault recorded for this batch · first left the corridor at 5% of the way through

the same as a picture (for printing)

What the chart shows. paa_flow carries most of it, running 36 sigma above the golden batch at 15% of the way through; 9 of 23 tags leave the corridor at some point; the first departure is at 5%, almost from the start.

The numbers. SPE 20597 against a limit of 520 (39.6x); 9.2% of the batch sits outside the golden corridor.

Driven by: paa_flow (41%), dumped_broth_flow (21%), ph (20%), acid_flow_rate (8%), base_flow_rate (3%)

23. Batch 70 - APC with Raman does not look like a golden batchseverity 10
APC with Raman · no fault recorded for this batch · first left the corridor at 5% of the way through

the same as a picture (for printing)

What the chart shows. paa_flow carries most of it, running 66 sigma below the golden batch at 38% of the way through; 7 of 23 tags leave the corridor at some point; the first departure is at 5%, almost from the start; only a small part of the batch is outside, so this is a short, sharp deviation rather than a batch that ran wrong throughout.

The numbers. SPE 19378 against a limit of 520 (37.3x); 2.7% of the batch sits outside the golden corridor.

Driven by: paa_flow (94%), acid_flow_rate (1%), temperature (1%), ph (1%), cool_water_flow (1%)

24. Batch 87 - APC with Raman does not look like a golden batchseverity 10
APC with Raman · no fault recorded for this batch · first left the corridor at 0% of the way through

the same as a picture (for printing)

What the chart shows. paa_flow carries most of it, running 32 sigma above the golden batch at 11% of the way through; 12 of 23 tags leave the corridor at some point; the first departure is at 0%, almost from the start; only a small part of the batch is outside, so this is a short, sharp deviation rather than a batch that ran wrong throughout.

The numbers. SPE 18538 against a limit of 520 (35.6x); 2.6% of the batch sits outside the golden corridor.

Driven by: paa_flow (51%), dumped_broth_flow (40%), sugar_feed_rate (2%), ph (1%), water_injection (1%)

25. Batch 69 - APC with Raman does not look like a golden batchseverity 10
APC with Raman · no fault recorded for this batch · first left the corridor at 10% of the way through

the same as a picture (for printing)

What the chart shows. paa_flow carries most of it, running 35 sigma above the golden batch at 11% of the way through; 10 of 23 tags leave the corridor at some point; the first departure is at 10%, early in the batch.

The numbers. SPE 18507 against a limit of 520 (35.6x); 4.2% of the batch sits outside the golden corridor.

Driven by: paa_flow (48%), dumped_broth_flow (36%), ph (6%), base_flow_rate (3%), sugar_feed_rate (2%)

5. How each way of running the plant behaves

Each curve is one batch: how far it sits from the golden batch at that point of its progress, taken across all 23 tags at once. The bold line is the middle batch of the group. All four charts share the same scale, so they can be read against each other.

Way of running it	Batches	Typical distance from golden	Worst batch	Mean yield	Beyond 10 sigma
recipe driven	30	0.82 sigma	152 sigma	2,912,750 kg	18 of 30
operator controlled	30	0.73 sigma	178 sigma	2,833,113 kg	8 of 30
APC with Raman	30	0.95 sigma	166 sigma	3,484,283 kg	30 of 30
faulty	10	1.20 sigma	183 sigma	2,600,203 kg	10 of 10

The four side by side

Only the middle batch of each group, so the four are directly comparable.

the same as a picture (for printing)

recipe driven — 30 batches

the same as a picture (for printing)

operator controlled — 30 batches

the same as a picture (for printing)

APC with Raman — 30 batches

the same as a picture (for printing)

faulty — 10 batches

the same as a picture (for printing)

6. Every batch, ranked

label	batch	strategy	band	max_z	sustained_z	worst_tag	worst_at_pct	outside_pct	first_outside_pct	T2	SPE	yield_total_kg
Batch 67 - APC with Raman	67	APC with Raman	far from it (over 10 sigma)	166	137.3	ph	47.5	20	2	25.36	3.054e+05	3.413e+06
Batch 91 - faulty	91	faulty	far from it (over 10 sigma)	155.5	144.7	substrate_conc	45.5	19.52	5.1	36.11	1.623e+05	2.597e+06
Batch 95 - faulty	95	faulty	far from it (over 10 sigma)	114.6	75.7	acid_flow_rate	62.6	12.78	10.1	20.92	4.98e+04	1.954e+06
Batch 99 - faulty	99	faulty	far from it (over 10 sigma)	134.9	95.9	acid_flow_rate	76.8	12.35	3	21	8.913e+04	8.908e+05
Batch 29 - recipe driven	29	recipe driven	far from it (over 10 sigma)	66.1	63.4	paa_flow	14.1	12	1	16.28	6.686e+04	1.842e+06
Batch 3 - recipe driven	3	recipe driven	far from it (over 10 sigma)	112.9	61.5	acid_flow_rate	98	11.96	1	17.02	5.879e+04	2.675e+06
Batch 75 - APC with Raman	75	APC with Raman	far from it (over 10 sigma)	61.8	51.5	heat_water_flow	14.1	10.61	2	15.48	3.138e+04	3.1e+06
Batch 5 - recipe driven	5	recipe driven	far from it (over 10 sigma)	66.7	66.1	head_pressure	21.2	10.35	2	14.12	8.248e+04	3.563e+06
Batch 15 - recipe driven	15	recipe driven	far from it (over 10 sigma)	66.7	65.4	head_pressure	21.2	10.35	2	11.76	6.912e+04	2.853e+06
Batch 53 - operator controlled	53	operator controlled	far from it (over 10 sigma)	52.3	45.9	water_injection	38.4	9.78	0	18.66	1.705e+04	2.858e+06
Batch 73 - APC with Raman	73	APC with Raman	far from it (over 10 sigma)	45.9	38.5	dumped_broth_flow	42.4	9.17	5.1	15.35	2.06e+04	3.549e+06
Batch 94 - faulty	94	faulty	far from it (over 10 sigma)	183.3	107.9	aeration_rate	46.5	7.17	9.1	33.99	1.088e+05	3.584e+06
Batch 61 - APC with Raman	61	APC with Raman	far from it (over 10 sigma)	45.9	38.9	dumped_broth_flow	54.5	6.39	3	11.63	1.785e+04	3.063e+06
Batch 45 - operator controlled	45	operator controlled	similar (2-5 sigma)	4.4	4.4	head_pressure	38.4	6.35	3	17.87	5.032	2.828e+06
Batch 17 - recipe driven	17	recipe driven	far from it (over 10 sigma)	65.4	53.5	water_injection	26.3	5.57	2	11.38	3.974e+04	3.852e+06
Batch 85 - APC with Raman	85	APC with Raman	far from it (over 10 sigma)	45.9	42.4	dumped_broth_flow	75.8	5.57	6.1	8.867	1.524e+04	4.044e+06
Batch 100 - faulty	100	faulty	far from it (over 10 sigma)	183.3	109.2	aeration_rate	46.5	5.52	9.1	18.9	1.049e+05	1.555e+06
Batch 97 - faulty	97	faulty	far from it (over 10 sigma)	162.4	50.8	acid_flow_rate	76.8	4.91	9.1	8.151	4.587e+04	3.894e+06
Batch 9 - recipe driven	9	recipe driven	far from it (over 10 sigma)	152.1	45.9	acid_flow_rate	60.6	4.57	10.1	10.17	4.727e+04	4.083e+06
Batch 35 - operator controlled	35	operator controlled	similar (2-5 sigma)	4.4	4.4	head_pressure	15.2	4.39	1	17.75	9.199	3.538e+06
Batch 23 - recipe driven	23	recipe driven	far from it (over 10 sigma)	56.1	39.1	paa_flow	15.2	4.17	3	7.662	1.089e+04	2.855e+06
Batch 69 - APC with Raman	69	APC with Raman	far from it (over 10 sigma)	45.9	42.7	dumped_broth_flow	58.6	4.17	10.1	3.815	1.851e+04	3.751e+06
Batch 89 - APC with Raman	89	APC with Raman	far from it (over 10 sigma)	38.1	31.9	paa_flow	39.4	3.65	5.1	8.008	1.549e+04	3.291e+06
Batch 93 - faulty	93	faulty	far from it (over 10 sigma)	135.2	23.5	temperature	42.4	3.65	1	15.13	2.317e+04	4.448e+06
Batch 66 - APC with Raman	66	APC with Raman	far from it (over 10 sigma)	32.4	27.8	paa_flow	12.1	3.48	3	7.151	1.058e+04	3.843e+06
Batch 83 - APC with Raman	83	APC with Raman	far from it (over 10 sigma)	45.9	38.1	dumped_broth_flow	54.5	3.26	5.1	7.593	1.742e+04	3.837e+06
Batch 77 - APC with Raman	77	APC with Raman	far from it (over 10 sigma)	45.9	37.7	dumped_broth_flow	49.5	3.26	10.1	4.764	1.685e+04	3.812e+06
Batch 7 - recipe driven	7	recipe driven	far from it (over 10 sigma)	74.9	65.2	ph	85.9	3.22	1	13.16	3.635e+04	3.455e+06
Batch 71 - APC with Raman	71	APC with Raman	far from it (over 10 sigma)	45.9	45.9	dumped_broth_flow	47.5	3.17	6.1	7.535	2.09e+04	2.831e+06
Batch 78 - APC with Raman	78	APC with Raman	far from it (over 10 sigma)	38.6	33.9	paa_flow	12.1	3.04	1	6.8	1.146e+04	3.589e+06
Batch 68 - APC with Raman	68	APC with Raman	far from it (over 10 sigma)	66.1	39.5	paa_flow	38.4	2.87	2	5.917	1.783e+04	3.573e+06
Batch 59 - operator controlled	59	operator controlled	far from it (over 10 sigma)	172	45.9	acid_flow_rate	84.8	2.87	7.1	8.357	4.227e+04	2.91e+06
Batch 70 - APC with Raman	70	APC with Raman	far from it (over 10 sigma)	66.1	42.6	paa_flow	38.4	2.7	5.1	6.664	1.938e+04	3.308e+06
Batch 44 - operator controlled	44	operator controlled	similar (2-5 sigma)	4.3	3.8	heat_water_flow	39.4	2.65	9.1	12.24	185.8	1.474e+06
Batch 87 - APC with Raman	87	APC with Raman	far from it (over 10 sigma)	45.9	42	dumped_broth_flow	45.5	2.61	0	9.026	1.854e+04	3.471e+06
Batch 6 - recipe driven	6	recipe driven	far from it (over 10 sigma)	14.5	8.7	acid_flow_rate	37.4	2.43	6.1	7.081	1945	2.389e+06
Batch 64 - APC with Raman	64	APC with Raman	far from it (over 10 sigma)	46.7	32.1	paa_flow	39.4	2.3	0	5.027	1.356e+04	3.362e+06
Batch 84 - APC with Raman	84	APC with Raman	far from it (over 10 sigma)	37.6	33.6	paa_flow	15.2	2.13	0	7.492	1.071e+04	3.488e+06
Batch 11 - recipe driven	11	recipe driven	far from it (over 10 sigma)	19.3	14.3	sugar_feed_rate	38.4	2.13	6.1	6.103	2385	1.965e+06
Batch 81 - APC with Raman	81	APC with Raman	far from it (over 10 sigma)	45.9	35.7	dumped_broth_flow	59.6	2.09	7.1	2.378	1.536e+04	3.186e+06

first 40 of 100 — the rest are in batch_scores.csv

7. The golden batch, tag by tag

The green band is where a good batch is expected to be at each point of its progress. The worst batches are drawn over it in red, and one reference batch in green for comparison. Drag to zoom, double-click to reset, click a name in the legend to hide it.

aeration_rate

the same as a picture (for printing)

sugar_feed_rate

the same as a picture (for printing)

acid_flow_rate

the same as a picture (for printing)

base_flow_rate

the same as a picture (for printing)

cool_water_flow

the same as a picture (for printing)

heat_water_flow

the same as a picture (for printing)

water_injection

the same as a picture (for printing)

head_pressure

the same as a picture (for printing)

dumped_broth_flow

the same as a picture (for printing)

substrate_conc

the same as a picture (for printing)

dissolved_o2

the same as a picture (for printing)

penicillin_conc

the same as a picture (for printing)

vessel_volume

the same as a picture (for printing)

vessel_weight

the same as a picture (for printing)

ph

the same as a picture (for printing)

temperature

the same as a picture (for printing)

generated_heat

the same as a picture (for printing)

co2_offgas

the same as a picture (for printing)

paa_flow

the same as a picture (for printing)

oil_flow

the same as a picture (for printing)

oxygen_uptake

the same as a picture (for printing)

o2_offgas

the same as a picture (for printing)

carbon_evolution

the same as a picture (for printing)

8. What this cannot tell you

Batches are aligned by percent complete. If a batch is slow because a phase ran long, alignment by phase or by an indicator variable would place it better.
The corridor is only as good as the reference batches. If a bad batch is in the reference set, its behaviour becomes normal.
The bands say how far a batch sat from the golden one. They do not say the deviation mattered: a batch inside the corridor throughout can still be poor if it is late, or if the yield is low for reasons no tag records.
"Worst tag" is where the deviation is largest, which is not always where it started. Use the per-batch charts above to see the order things moved in.

9. Comparing the ways of running the plant

APC with Raman produces the most and is also the most repeatable.
Across all 3 strategies the yields differ: one-way ANOVA F=9.91, p=1.3e-04; Kruskal-Wallis p=5.1e-04 without assuming normality. The spreads differ too: Levene p=5.6e-05.
The 10 faulty batches are excluded from every comparison below. They appear only in the last table, as the thing each strategy's golden batch is asked to detect.

9.1 What each way of running it produced

Way of running it	Batches	Mean yield (kg)	Spread (CV)	Worst batch	Best batch	Mean hours	kg per hour
recipe driven	30	2,912,750	26.4%	1,842,400	4,083,100	228.8	12,917
operator controlled	30	2,833,113	24.0%	1,474,100	4,196,000	229.2	12,431
APC with Raman	30	3,484,283	8.8%	2,830,800	4,044,300	224.7	15,587

CV is the batch-to-batch spread as a percentage of the mean. In a regulated plant it often matters more than the mean, because it is what lets you promise a delivery date and pass a validation.

9.2 Is the difference real, or is it noise?

Comparison	Difference		Welch p	Mann-Whitney p	Cohen d	Significant
recipe driven vs operator controlled	+79,637 kg	+2.8%	6.7e-01	8.9e-01	+0.11	no
recipe driven vs APC with Raman	-571,533 kg	-16.4%	5.3e-04	1.1e-02	-0.98	yes
operator controlled vs APC with Raman	-651,170 kg	-18.7%	2.3e-05	4.4e-05	-1.23	yes

Two tests because one assumes normality and the other does not; they should agree. Cohen d is the size of the gap in standard deviations: 0.2 is small, 0.5 medium, 0.8 large. A difference can be statistically significant and still too small to act on, so read the size as well as the p-value.

9.3 The golden batch of each way of running it

One real batch per strategy, chosen on two criteria at once: high yield and close to that group's own mean trajectory. The highest-yielding batch is often a lucky outlier that ran unlike anything else, and the most typical batch is often mediocre. These are both.

Way of running it	Golden batch	Yield	Yield rank	Closeness rank
recipe driven	Batch 26 - recipe driven	3,886,800 kg	4 of 30	2 of 30
operator controlled	Batch 60 - operator controlled	3,802,300 kg	3 of 30	5 of 30
APC with Raman	Batch 79 - APC with Raman	3,825,600 kg	5 of 30	2 of 30

9.4 Does each way of running it actually look different?

A golden batch is built from one strategy, holding back some of its own batches, and every group is then scored against it. The figure is the median batch's typical distance, in sigma. The shaded cell on the diagonal is that strategy's own held-back batches — the fair self-comparison, and it should be the lowest number in its row.

Golden built from	recipe driven	operator controlled	APC with Raman	faulty
recipe driven	0.77	1.07	1.08	1.11
operator controlled	0.82	0.74	0.95	1.20
APC with Raman	1.05	1.35	0.66	12.51

Golden built from	Its own held-back batches	The other strategies	How much further	Fault detection power
recipe driven	0.77	1.07	1.40x	1.4x
operator controlled	0.74	0.89	1.20x	1.6x
APC with Raman	0.66	1.20	1.82x	19.0x

"How much further" says how distinctive that way of running the plant is. "Fault detection power" is how far the faulty batches sit compared with its own good ones — a tighter normal is a more sensitive detector, which is a reason to run a strategy quite separate from what it yields.

9.5 What this comparison cannot settle

Do not rank strategies by distance from the main golden batch in section 2. That golden is built from all the good batches, so it is a blend of every strategy, and each is being measured against an average that contains itself. Section 9.4 avoids that by giving each strategy its own golden.
These are the batches that were run, not a designed experiment. If the strategies were used in different periods, on different feedstock or by different crews, that difference is inside the numbers and cannot be separated out.
Yield is the only outcome recorded here. Cost, energy, operator effort and the price of the instrumentation are not, and a strategy that yields more can still lose on all four.
Sample size is 30 to 30 batches per strategy. Enough to see a large difference, not enough to see a small one.

Folder: /Volumes/YN-4T/WIP/CTRL-Designer/B-CDL-Data-Analysis/E-Batch/IndPenSim-Golden-Operator