Clarification on model ACEc and ACE

[jinningwang]

#585

• Fix ACEc.bias.tex_name
• Change ACEc.bias.default to -1 to follow NERC's definition
• Add references

[jinningwang]

@cuihantao
One question regarding this PR.

As here I changed the default value and equation, it seems to be a potential breaking change. Before manually updating parameters of ACEc.bias, users can run into unexpected results.

So, I'd like your suggesstion that which option would be better for us?

1. Revise the equations: use abs() to eliminate the impact of sign
2. If there is a parameter correction step, correct positive values to negative

[cuihantao]

If it's always going to be a negative value, you can use "abs". This is to safeguard users who entered the value mistakenly as a positive value.

If both positive and negative values have a meaning, then we should do any correction to parameter but rather correct the equation.

[jinningwang]

Bias and frequency response are both expressed as negative numbers. In other words, as frequency drops, MW output (β) or desired output (B) increases. Both are measured in MW/0.1 Hz

Given NERC's suggestion, I update the PR with first option.

[jinningwang]

@cuihantao ready to merge

[cuihantao]

Jinning, thanks!! I'll tag a new version shortly.